Aasemoon'z Twine Aasemoon'z Twine / Items

A vision lost

Get Feed
A vision lost
Description
Twine - A vision lost

Motivation

This post is written in the agony of realization. The painful realization that months and months of work may possibly have been a wasted effort on my part - as well as the part of others.

The story begins…

The Promised Land

I do not recall where the name first came up. And it wasn’t immediately clear to me what the significance of the name was. Regardless of how or why, I somehow found myself on a mysterious web page for a new startup company touting in vague terms a social web service like no other. It was Twine.

I remember seeing a good looking front page, linking to a beta candidate signup form. Scanning the description of the service, I found several buzzwords that had been on my radar for years. Including RDF, RDFS, OWL, SPARQL and friends. All good stuff from the realm of Knowledge Representation (KR). The ethereal plane from where truth, or at least probabalistic truth, could be found. A place where all knowledge is broken down into statements, also known as triples, forming graphs of interconnected nodes in a uniform and truly scalable way - scalable in the sense of compositionality.

This (KR) was the kind of stuff I had studied on my own while specializing my education. As with programming language concepts, data concepts were just as fascinating. With KR data comes with it’s own semantics and a so-called inference engine takes the higher facts and uses these to form new facts. One example is transitivity; say there two facts in the knowledge base: a contains b and b contains c (well actually there are more facts, like the fact that contains is a transitive relation), then it is possible to infer that a contains c. This is a ...

Original URL
http://xosfaere.wordpress.com/2009/02/27/twine-a-vision-lost/

Comments

  • Public Comments

    • Aasemoon . Aasemoon .
      9 months ago


      Self Explanatory!
      Aasemoon'z Twine

    • Kurt Laitner Kurt Laitner
      9 months ago


      careful! you might get twained
      Aasemoon'z Twine

      • Aasemoon . Aasemoon .
        9 months ago


        Or I may very well twain myself! =P
        Aasemoon'z Twine

      • 9 months ago


        What happened to Twain? I took a hiatus myself for awhile for similar reasons to Bent's but came back to find the most useful things missing and all the new stuff just eye candy (for the most part). Seems like everyone is bailing (again, for the most part).
        Aasemoon'z Twine

        • Kurt Laitner Kurt Laitner
          9 months ago


          you have been away...
          Aasemoon'z Twine

        • . .•∞º˙ . .•∞º˙
          9 months ago


          @Ryan-

          Twain was initially kicked out of Twine for being too vociferous on the issue of shutting down Twine User Feedback And Suggestions (formerly Beta Feedback and Suggestions). RN had decided to close the ability to comment on that Twine because of repeated flame-outs between parties, and because (apparently) 'other' users were complaining that their Interest Feed was receiving too much jabber from that Twine that they didn't want to read. (Didn't want to turn it off though, like a good soap opera!)

          Anyway, Twain's actions in response to the closing of commentary on TUFAS became very vocal, and RN decided to pull the plug and cut her off. As part of that action, they also DELETED all of the content, commentary, and contributions that Twain had built up over the course of 8 months or so of activity within the system. See for yourself. Try this search:
          http://www.twine.com/search?type=Member&text=twain

          or even this one:

          http://www.twine.com/search?type=&text=twain

          Many of the Twines that she created are still around, and still very active--they simply show no-one as the owner. All of her comments are gone. Replies to her comments are orphaned. Entire threads have been rendered meaningless because her interactivity no longer exists on the system. She has been Big Brothered. She has been Twained. She is an ex-twiner.

          I say initially, because several members successfully lobbied for her return. So RN allowed her to re-enroll, but without her previous account, and without any of the previous items that she had contributed. She even had to rebuild all her connections again from scratch. (Which because she is Twain, only took a few days.) She was put on "probation" at the time with a warning about any further outbursts, and things progressed smoothly for a time. But then issues arose again surrounding the flame-out problems that had taken place within TUFAS, and Twain has now been deleted for good.

          The lesson has become "Beware of being an outspoken critic because the authorities may decide to make you irrelevant", much like the Chinese government has been attempting to do with Tibet for many years...
          Aasemoon'z Twine

          • 9 months ago


            That sucks. Reminds me of Mandel from The WELL. I'm quite surprised anyone stuck around after that. Of course, I tend to be a bit "idealistic" when it comes to such things... :) I more fully understand Bent's and everyone else's responses now.
            Aasemoon'z Twine

          • Scott White
            9 months ago


            Without making any judgements on Twain and the whole situation, since it is a highly sensitivie topic, it is my understanding that she was evicted only after violating site policies and as a result of that her content was removed although not deleted from the db. Although I do not work in customer service, but I have never heard of another user being evicted other than ones posting porn, etc.
            Aasemoon'z Twine

            • . .•∞º˙ . .•∞º˙
              9 months ago


              Yes, essentially we agree. The downside is that the removal of her content also either orphaned or completely eliminated any associated content--comments, posts, bookmarks--that were embedded within items she had contributed. So not only is her deletion a loss of a member, it is also a loss of a great deal of collateral content that is no longer accessible. I personally lost almost 40% of the content I had contributed to a private collaborative Twine that we had been working on while attempting to develop a gaming system. What's left has been rendered gibberish, with little hope of resuscitation at this stage. 4 months of work is now irretrievable.

              Twain's actions were her own, and it is not my intent to defend her. I was not involved in her eradication, yet I (and many other members) were "punished" just as hard. The whole event has left a very bad taste in my mouth. The lesson learned has become "beware of what you contribute, as it may not be here tomorrow."

              If as you say, her content still resides in the DB, then perhaps all is not lost, but I'm fully aware at this stage that the needs of one user carry far less value than building the system. I do however feel that members need to be aware of the risks they run when using the system. And in the long run, I hope that an issue like this does not arise again.

              I wish I had Bent's skills at programming--I would have built an off-line archiving mechanism many many moons ago...alas, back to my crayons.
              Aasemoon'z Twine

              • Scott White
                9 months ago


                I was not aware of this. I will talk to powers that be and see what can be done. This certainly was not intended and hopefully will never happen again. I can totally understand your frustration.
                Aasemoon'z Twine

      • . .•∞º˙ . .•∞º˙
        9 months ago


        Is THAT a Train Wreck??? BWaahahhaaaaaaaa
        Aasemoon'z Twine

    • johnrod johnrod
      9 months ago


      Fine example of the writer's craft.
      Aasemoon'z Twine

    • . .•∞º˙ . .•∞º˙
      9 months ago


      and sadly, also HIGHLY ACCURATE.
      Aasemoon'z Twine

    • Bent Rasmussen Bent Rasmussen
      9 months ago


      Thanks people. I am willing to risk my existence on Twine for this matter. It can't get much worse anyway...
      Aasemoon'z Twine

      • 9 months ago


        So what are your thoughts on an OSS alternative? Twine's promise beyond the walled garden was its as-yet-still-missing API's. I imagine many of us would be willing to pitch in on the development were it open. Alas, it's not. You had at one point started thinking on (and developing?) an application to publish items into Twine. Whatever happened to that idea (beyond the lack of connectivity into Twine)?
        Aasemoon'z Twine

        • Bent Rasmussen Bent Rasmussen
          9 months ago


          It is possible to auto-submit data into Twine via HTTP POST. I just never got further into it, somehow when I get home from work I don't feel like programming more.
          At this moment (started yesterday), I'm working on a console application to fully back up all my data. It will be a simple piece of code, but it will secure the data.
          I can share the code with you soon... Basically just an ultra simple web crawler targeted at Twine and at one user account. I can make it lazy so it wohn't stress RN's servers.
          Aasemoon'z Twine

          • Kurt Laitner Kurt Laitner
            9 months ago


            how much would it cost us to have you stay home for a day?
            Aasemoon'z Twine

            • Bent Rasmussen Bent Rasmussen
              9 months ago


              Ryan has a quick pre-alpha snapshot. It's very simple recursive code. Just takes a couple of hours to figure out the new .Net APIs around syndication, etc. But it basically decomposes from a user name to a search feed to a set of twines to a set of items. Need some debugging time tomorrow. I usually don't program outside work...
              Aasemoon'z Twine

              • Kurt Laitner Kurt Laitner
                9 months ago


                yer just supposed to drink when you program outside of work, then it can't be work, cause they wouldn't let you drink.. see?... oh yeah this is the internet
                Aasemoon'z Twine

              • 9 months ago


                The initial snapshot looked good. Last night was date night with my wife, but I'll be taking a deeper look over lunch and tonight. I'm happy to help debug and, if possible, add functionality.
                Aasemoon'z Twine

          • James Todd James Todd
            9 months ago


            people have built firefox extensions that bookmark into twine. pretty easy actually.
            Aasemoon'z Twine

          • Scott White
            9 months ago


            I hope you're using the RDF exporting capabilities we provide. It wil make it a lot easier. Of course, once the API is released that should suit your needs perfectly. But I know an API that is implemented but has not been released doesn't do anyone any good :)
            Aasemoon'z Twine

            • Calvin Smith Calvin Smith
              9 months ago


              Unfortunately, the most desirable stuff to export is not available in RDF. For example, the /my-items page is not available in RDF. If you try to navigate to /my-items?rdf via a browser or to access it programmatically with an "Accept: application/rdf+xml" header, you'll get an HTTP 406 (NOT_ACCEPTABLE) error.
              Aasemoon'z Twine

              • James Todd James Todd
                9 months ago


                i believe each feed has an alternate link to the entities alternate representations (be it html, rdf, etc). so, one could paginate through feeds (collections) and incrementally retrieve each feed entry's rdf representation?

                for today at least. things are and will be improving.
                Aasemoon'z Twine

            • Bent Rasmussen Bent Rasmussen
              9 months ago


              Hi Scott, Indeed I am using the RDF export, but the RDF export is only partial, especially for items.
              Aasemoon'z Twine

        • Nova Spivack Nova Spivack
          9 months ago


          Maybe you'll all get to pitch in on our own OSS platform version. If we can ever get it ready for release. It's kind of a free-time project right now, but we're hoping it will make it out the door this year. We'll need a small army of Java / Semantic Web experts.
          Aasemoon'z Twine

        • Scott White
          9 months ago


          API is actually already there. It has not been turned on yet because performance issues are still being worked on.
          Aasemoon'z Twine

          • Bent Rasmussen Bent Rasmussen
            9 months ago


            "Just" REST is more than enough :-)
            Incl. complete RDF export for items. I'd also like to build my own presentation templates since I am really hurting watching the bland list view of things...
            Aasemoon'z Twine

    • Greg Boutin Greg Boutin
      9 months ago


      Thanks for publicly expressing yourself Aasemoon. With all the unconstrained personal attack trash I got from Nova for my article http://www.semanticsincorporated.com/2009/02/as-promised-interview-with-twine-on-the-usability-question.html, I must unfortunately say that your defection comes as a breath of fresh air... I don't think I would ever have wished for that, but he made it happen and personally, I know where all the Twine problems take their roots in.
      Aasemoon'z Twine

      • Aasemoon . Aasemoon .
        9 months ago


        Hi Greg... thanks for your comment, it's very interesting to read your article. The author of the article that I just shared here has in fact been one of the most active contributes to Twine since the very early days, and his article is doing a very nice job summing up everything that many of us have been feeling during the past months, and specially now, which is why I decided to share this. It's voicing the frustration and concern growing in all of us. So yes, I agree that it's a breath of fresh air!
        Aasemoon'z Twine

      • . .•∞º˙ . .•∞º˙
        9 months ago


        Yes, Greg, you've experienced first hand the magician's trick of waving the magic wand over here, while the cards are stacked over there. Rather eye opening experience, isn't it? It also makes some of the comments on Venture Beat take on a slightly different perspective.

        Great blog btw. It's always good to validate one's own experiences with that of others.

        The bigger issue is I think, we all really enjoy using Twine. Which is why everyone is so upset about the things that are and are not taking place. Bent's blog REALLY sums that all up very eloquently. And I truly hope that SOMEONE decides to listen for a change.

        (@bent--I'm on the gangplank right behind ya...)
        Aasemoon'z Twine

    • Nova Spivack Nova Spivack
      9 months ago


      Hi folks, I'm saddened by this thread since many of you are among my favorite long-time Twine beta-testers who were there from the beginning. I hope this reply will be the beginning of a path to addressing some of the concerns you are voicing.

      I want you to know we have been listening and are working on most of the things you are asking for. It's just that these things take time, and I think expectations may be a bit higher than is realistic as well sometimes. But nonetheless, there is also a lot of valid concern and many of the requests in Bent's blog post have a lot of merit (and in fact several are being addressed in the release this week, ironically).

      So first of all, I'm sorry to those of you who don't feel listened to as much as in the old days. Yes, things are different now -- Twine is a lot bigger and there are a lot more users and we get a lot more feedback (via Support), and so there has not been as much direct interaction with users from me, for example. I wish I had more time to interact with users more often.

      But, I want to reassure you that we (and I) really do listen to the feedback we get through many different channels -- especially The Lounge, as well as all the emails and other feedback we get to support and directly as well. I get regular reports on what the top requests and issues are from the Lounge, and we also do random sample surveys of the broader user population, plus a ton of measurement of what features are getting used or not.

      Since there really is only a small team working on the community and support side of things (1 support person + 1 community person) and there is an extremely large number of users (not just members of the Lounge, but also hundreds of thousands of others) it is hard to give a personal response to every message. But that doesn't mean the team isn't paying attention.

      In any case, for the members of the Lounge -- who were in the original group of beta testers -- they are special (because they have been involved for so long and because they have more experience using Twine than most new users) and so we created the Lounge in order to interact more closely with them.The idea was to make the Lounge private so we could have discussions that were not meant for newbies, and so we could rollout early access to new features there for example. This was supposed to be a special private area for some of our most active users to give us feedback and to try new things out that were not ready for general release. In any case, I will delve into our reasons for doing this a little bit more below.

      But first let me acknowledge the pain I'm hearing from some of you. I know that you are among our most passionate users and when things frustrate you, you feel that more than others. While collectively those who we have in the past call "power users" represent a small fraction of the total audience today, you are among the most active users and you tend to push Twine to (and beyond) the limits of the system before general users. Often your needs are a bit higher end than the average user of Twine, and that is why we look to you for advice and first-reports of problems, bugs, or improvements we need.

      The second issue I want to acknowledge is that Twine has not yet lived up to its full potential. But I really believe it will. It's just going to take time. For many reasons. That said, I think you will see some positive progress soon, and for the rest of the year. However it probably won't be advanced knowledge representation or AI in the near-term. We have some of that in the backend, but getting it to scale to this kind of audience, and making it understandable to the average consumer are daunting challenges. And so we have opted to first focus on the basics to make sure that Twine can (a) scale, and (b) solve some real problems for average users.

      Below I'll try to provide some insight into where things are right now, since much of what I'm hearing above is that you want more visibility into where we are going together in the coming year. This year we are working on many improvements which I think will speak to many, but not all, of the requests echoed in Bent's original post. I'll outline some of them below.

      First of all, we release a new build every 3 weeks. That limits how much can be released per build because it is only 3 weeks after all. Some features take months to build and this is simply normal. Many other services release once every 6 months. We opted for every 3 weeks. It's a tradeoff. If we release every 3 weeks, the releases will be smaller, but over 6 month timeframes, you will see the same amount of progress, or more, than sites that release every 6 months. That is something to keep in mind. However, it doesn't address questions like why haven't we launched next-generation OWL-DL capabilities? There are some features which are longed for by people who know a lot about the underlying Semantic Web standards -- such as reasoning, more powerful uses of ontologies, etc. -- that simply are very very difficult to roll out on a site this large, or for a consumer audience. We have built the engine to do this over time, but they are going to be small steps, and it will take a while. Maybe some of these steps will be too small for some of you who were hoping for these sooner. I certainly understand.

      In any case, aside for advanced Semantic Web reasoning, etc. You WILL see several big changes and improvements this year. Including in the release that is going out this week actually.

      This week, for example, several of the wished-for improvements on Bent's list are going out:

      1. The new Advanced Bookmarklet will become the default for all users. It will failover gracefully to the Basic Bookmarklet for those pages that it does not work on (which are few, we believe). So this should be a big improvement.

      2. We are soft-launching a new version of the Interest Feed that is going to be awesome. We'll test it out in this release and then hard-launch it for all users in the next release (3 weeks from now) assuming there are no major issues that need to be addressed first. You will be able to try out this new version via a URL that we'll be sharing with members of the Lounge.

      3. In addition, we are starting to roll out more powerful moderation capabilities (a whole bunch of them).

      There are also many other UI fixes going out that are less major in this release.

      Beyond this week however, in coming months we will be improving a number of sections of the site, one by one. Each one takes time and we do a lot of testing so they won't happen all at once. But here is what we are actively working on (in various stages of planning, design, engineering, testing):

      - Revamping the "Item Detail" page design
      - Redesigning the "My Items" functionality
      - Adding in ads and beginning to test how to best do that
      - Adding in new algorithms that show what is "Most Popular" in Twine as a whole as well as within individual twines.
      - Improvements to recommendations (based on the above)
      - Better integration with Twitter (enabling you to tweet out what you add to Twine to your Twitter account, for example)
      - We are also working on improved design for the Global Navigation
      - We are midway through development of new Visualization features to visualize what is going on in Twine in some new ways. That will launch in a few months.
      - We have an Alpha of our API in early testing, enabling access to the semantic data within Twine so folks can write their own clients. We hope to release a more powerful version later this year.
      - Search is also an area that could benefit from further improvements and redesign and we're working on that.
      - We are doing a major architecture upgrade to speed up the site and enable more efficient handling of data. This is already in process but will take many months to complete as it is a big backend project and will happen in stages.
      - Social network integration is underway to enable sharing of data (both in and out) from other social networks (Twitter is first, others will follow...).

      As you can see, there is a lot going on, and it's a small team. We're doing the work of a much larger company actually, and the price we pay is that we can't work on every feature at the same time -- we focus on different areas for each release.

      There is more than the list above that we are working on as well. These fall into the "We'll get to it if we have time" category this year: Particularly in the areas of surfacing more advanced semantics in the service. Although all the data is actually stored semantically, we fully realize we have not exposed much of the semantics (or the full potential value of the semantics) in the site yet. Some of this will be solved when we open up an API. But I personally think it will probably be a while before we fully live up to my own (and other people's) vision for the site. What I would like to see happen eventually is for Twine to do more machine learning, automatic organization of data, auto-population of twines by learning about their content from users, and in particular more use of RDF both in terms of the data we put out and the data we might mine in. I would like for example to see us make the bookmarklet extensible so that it can recognize RDFa from other sites and pull it in appropriately. I also think that we could do a lot more in our search feature to search on specific properties of things in Twine. And we could add more types to our ontology (like reviews, recipes, etc.) -- although we don't yet have much evidence that users want more types (let us know if you do, and which ones you would want!).

      As for the question I've had regarding whether or not we should do rollbacks of features that are found to have bugs – sometimes rollbacks are worse than not rolling back (and they take up to an entire day or more to test and push new code, which diverts the dev team and forces the next rev to be late). The dev team generally does not like rolling back because of this, unless it is a major emergency. In the case of the Twitter bug a few weeks ago, since a fix was only a few days away, rolling back was decided against because it would have taken just as long as a fix and would have actually delayed the fix in the end. So it’s not always an easy choice. Sometimes there is only a less-bad option. If we just didn’t release every 3 weeks, and instead did a release every 6 months like many other companies these problems would not happen, but then people wouldn’t get to be as involved in the development. You see the challenge.

      As for why we closed the public Feedback Twine and opted for a private Lounge twine instead for that kind of feedback -- it was because we wanted to work more closely with the early beta testers and we wanted to be able to pre-release and alpha-test features there, privately, with a core group of Twine pro's rather than the general user population. The Feedback Twine was starting to get a lot of comments from new users who really were looking for support or who were asking questions that had already been asked and/or answered many times before -- we wanted to direct the more support-oriented requests to our new Support ticketing system, which is much more manageable for large numbers of users than a threaded discussion as I'm sure you would agree. So we set up the Lounge instead for the original beta testers to participate in and directed new users to Support.

      I also want to add that we have no problem with people giving us constructive feedback, and we have never kicked anyone out for doing that. One person in particular WAS kicked out, but not for giving us feedback -- it was for breaking the site policies. In all fairness, we did let that user back in with a warning, but the problems continued. Unfortunately when we finally did kick them out a second time, it caused some of their content (much of which was actually quite good content) to get deleted and that was a loss. It was still early in the beta and we simply didn't have tools to salvage all their content. We do have a better moderation and account management system today. That was unfortunate for everyone -- all of us -- because we all really liked that user on their positive days. These things happen in online communities and I wish we all had been able to avoid that unfortunate incident.

      Finally -- why did we remove the New Twines page? It was because a small group of people were using it to spam Twine with new twines that were basically just ads or worse. That was causing a lot of work on our end to constantly delete those twines, and it was detracting from the value of the service for everyone. Now we have much more powerful automated spam detection and removal scripts running which have resolved much of that issue. But I really think that the New Twines page needs a rethink and certainly needs to be moderated. We've given that some thought and will be bringing something in to help people discover new twines, and the long tail of twines, in the future. Personally, I wonder if there is really any value in seeing brand new but empty twines anyway. I think it would be much better to only show New Twines that have a certain amount of usage first as a minimum threshold. That is probably what should happen if we bring that page back. Feel free to give suggestions in the Lounge or directly to support -- those DO go to the product team!

      I also have been advocating for a voting system in Twine, so that folks can vote on items they like or don't like. As some of you recall we launched an early attempt but users were confused by it, and actually it did not get much usage, so we took it out and decided it needed more thought on our end. We do plan to bring it back in later this year. However we have also found that with some simple algorithms we can also derive what people like best automatically. So what is better: Asking people to vote manually or just learning automatically? Or both?

      As for site performance for holders of very large accounts (hundreds to tens of thousands of items added, and hundreds to thousands of social connections), this is a known-issue and only occurs for a very small number of users who happen to have such large accounts (I'm one of them). It is indeed very frustrating and is basically due to the fact that these extra-large accounts are simply pushing Twine past the limits of the beta-architecture. To solve that we are putting in the next-evolution of the backend and that will greatly improve performance overall, and certainly for large account users (like me). It will take several more months however and I know it is hard to wait for that. Meanwhile, since I am in that category too, you can be sure I am on top of that. So, yes, that is going to be fixed. But also keep in mind that this concern, is something applies to less than 1% of the user base today, so in terms of priority we do have to also make sure we are focusing on the other 99.9% -- who simply want the global nav to make more sense, etc.

      Hope this helps to provide more visibility to those of you who are wondering where Twine is actually headed.

      We are working hard to improve the service and I think you will be impressed with progress this year.

      Very sorry it hasn't been fast enough for some of you. With a service like this, with so many moving parts, the reality is that as it gets larger (and as we get more users), things do take longer to change (and we get a lot more feedback too). In the past the rate of change was faster because it was a very new service. Now we are trying to be a lot more thoughtful about every new feature and change. That's (probably?) a good thing.
      Aasemoon'z Twine

      • 9 months ago


        @Nova: Thanks for taking the time to respond here. I'm sure a lot of this sort of banter lately must strike a bit painfully and even perhaps seem unkind. I hear you saying many of the things I think we all want to hear; it's just hard to have used a lot of the features that have since been removed that were very core to our early excitement.

        Having recently come back after some of the first feature removals, I was quite surprised at the changes, both for good and bad. The bad struck me the most. After deciding to try Twine out again, I convinced some co-workers to join me here and described my earlier experience, which was mostly positive. We were all a little disappointed in what we found, especially the lack of rich, semantic editing available on each item. For the most part, we've resorted to using Twine as a way of sharing bookmarks, though only because none of my other co-workers use any other bookmarking system.

        Twine certainly holds lots of promise, and I'm glad to know that it has someone as passionate about its promise backing and driving it. However, by its very nature Twine does and will primarily attract power users. I hope you don't allow Twine's potential to be watered down by the additional promise of more users given a simpler system. The idealist in me believes that if you build a great tool, the users will come. I realize that's got to be tempered with reality to a degree, but the principle holds. Build a great system for your core users, even if they begin few in number, and let them build your user base.
        Aasemoon'z Twine

        • Kurt Laitner Kurt Laitner
          9 months ago


          Excellent point, well covered in the book 'Crossing the Chasm' and in this case the core users are in fact your beachhead market, where you can be a big fish in a small pond and create reference-able clients. Given scaling issues it may not even be a good idea to pursue hockey stick growth, even if it gets the investors and M&A prospects excited. For semantic functionality to become mainstream and 'usable' by the average user, the advanced users must first experiment playfully with the tools and create the first working machines with the new parts, iterate until patterns become apparent, then those patterns can be cleaned up and served to the unwashed masses, with uptake being carefully noted and commercial properties built around them. I do not think front end of the semantic web will be designed then delivered, it will be discovered in use.
          Aasemoon'z Twine

          • Nova Spivack Nova Spivack
            9 months ago


            I hear you Kurt. At the same time we can't stay in beta forever. In order to really achieve our goals (which will take several years) we have to build a sustainable revenue stream, which means attracting mainstream users and making money from ads (unless you have a better idea, ha).
            Aasemoon'z Twine

            • Kurt Laitner Kurt Laitner
              9 months ago


              I might, but you'd have to pay for that :-)
              Aasemoon'z Twine

              • Nova Spivack Nova Spivack
                9 months ago


                I would be glad to :)
                Aasemoon'z Twine

                • . .•∞º˙ . .•∞º˙
                  9 months ago


                  @nova--thanks for filling us in. I think your responses go a long way towards enlightening us all to current and future status of Twine. I for one feel better now...
                  Aasemoon'z Twine

                  • Nova Spivack Nova Spivack
                    9 months ago


                    I appreciate that. I am happy to provide this kind of detail. You all have put so much time into Twine, you should definitely have visibility into where we are, where we are going, and why we are making the tradeoffs we are making right now in prioritizing features around usability first.
                    Aasemoon'z Twine

                • Kurt Laitner Kurt Laitner
                  9 months ago


                  I'm sure you've got more than ads to eyeballs in your monetization strategy already, so send me a list and I'll know if you are serious. happy to sign an NDA
                  Aasemoon'z Twine

                  • Nova Spivack Nova Spivack
                    9 months ago


                    Actually we really are starting with just ads for now. We've considered subscriptions for pro accounts, pro-private twines, enterprise accounts, and many other ideas. But starting with ads is simplest to begin with. Our recommendation algorithm has lots of potential for better targeting of ads in the future too, but we'll probably begin by just testing simple Google ads for example. We have to prove that we get better than average clickthroughs due to the passionate, vertical interest nature of Twine first. If we can prove that then it makes sense to put more effort into ads. If we can't prove that, then we may look at other ways to bring in revenues. I believe however that we must keep Twine free for consumer users, and for that advertising is a great time-tested way to monetize. However it depends on large traffic volumes and a highly active consumer audience to work. Time will tell as we test these ideas out. I do like your suggestion of NDA's for a panel of Twine member-advisors who want to be more involved in our thinking on this. Will mention to Bonnie.
                    Aasemoon'z Twine

        • Nova Spivack Nova Spivack
          9 months ago


          I am particularly interested to hear about what features were removed that are of great value. It may actually be the case -- at least for some of them -- that we are adding them back in in the future, but with a better design and improved functionality. Certainly feel free to let me know.

          I can guarantee everyone here that my goals for Twine are MORE ambitious than any of what I've heard from anyone. And if there is anyone who at times feels frustrated that those goals have not yet been achieved, it is me. Because I truly know what the potential of these technologies is, and Twine is tantalizingly close. But not there yet. Knowing what I know about the power of our backend, and the talent on our team, it is all the more difficult to delay some of the more exciting and advanced features in favor of some of the more basic and less-exciting features. The team often has to remind me, when I get impatient about such features too, that after spending a year focusing exclusively on the needs of beta-users and power-users, we now do need to catch up and focus on the needs of the general population for a while. That has been the goal for the first two quarters of this year. There is much, much improvement needed to the basic functionality and design of the site and the team really knows this because they hear it every day from new users. That is really where the team feels the most people have the most pain, and in all fairness, that is correct, and I support the decision to focus on new users and less advanced users for a bit. That will result in Twine being better for everyone in the end.

          At the same time, those of us who know more about RDF, OWL, SPARQL, etc, are seeking more. And actually... we have something in the lab that addresses folks who want to really get in there and tinker with the semantics themselves, build their own apps, and do things that are just not possible within the limitations of Twine's interface today. We're thinking about the right way to release that. It's not ready yet but could be in about 6 - 9 months. No promises because it's really a question of resources and focus for our small team -- but we would love to find a way to release some of that more advanced functionality somehow.

          Meanwhile, our goal for Twine is to really prepare the service for larger adoption by a more mainstream audience. The fact is that if you look at our stats, we are seeing very rapid growth and adoption by a very different type of user than before. In the days of the Beta test, Twine users were a particular breed. Power Users. Today, the vast, vast, vast majority of users are totally new to all of this and don't know what "semantic" means, and they don't care. IF we can succeed in making Twine something they can really understand and benefit from, this will benefit everyone. Because if Twine really takes off (and it is too early to tell right now what will happen yet) then we will be able to eventually achieve the more ambitious goals we have for the site. We're going to need a few years of runway for that to happen though -- and so we have to make Twine something that general mainstream users love and understand.

          But I get your point about striking the right balance. It is something we struggle with every day and we spend countless hours in development and product team meetings debating. Tradeoffs of this nature are always hard. Like you, I am an idealist too. But we have to be realists in this situation. The vast majority of users have very different, and much simpler, needs than we do. And they have much less tolerance for anything remotely techie or beta. Such is the challenge we face in figuring out what features to prioritize. That said, I have tremendous faith in our product and engineering teams. They are just awesome. I truly have faith that with time we will get there.

          I know some of you may want more than we can deliver in the timeframe we are able to work in. I can only say I understand. But if you look at the rest of the Semantic apps out there -- you'll see that we are all in the same boat. There's a lot more work to be done. At least however, Twine is growing faster and getting more adoption than the rest of the pack right now. For us at least that is a sign we must be doing something right. Perhaps some of you won't be willing to wait while we work on the basics. I invite you to come back later on and give us a second chance when we are farther along. For those of you who want to come along for the ride, we're happy to have you with us.
          Aasemoon'z Twine

          • Kurt Laitner Kurt Laitner
            9 months ago


            There are several levels of usability.

            First there is the ability to get simple things done simply. System performance and reliability feature heavily at this level. Adding, removing, sharing, discussing, seeing what is new, seeing what you care about most, and twine's bread and butter, discovering new things when you get bored. Stability is a challenge currently, performance is a challenge, for some the system is broken - hopefully the architectural improvements will get this right over the next few months.

            The next level would be to be able to create complex structures out of simple parts, in this way advanced users can expand the usage of the system and others can follow the paths they blaze. For example, the ability to feed one twine into another, to twine twine pages and other twines with behavior controls would allow new structures to be built, and their uptake studied. Another example is the building of user types, and the interaction of those types with other entities. Another is a powerful search/filter system that allows advanced users to create tight streams for less advanced users to consume. Combining twining of twine content with search/filtering is very powerful. These are not things an average person would ask for, but they will enjoy the result of their use. Empower curators to create the exhibits for the masses.

            Finally, is to take the complex structures that have resulted and simplify their expression for those who cannot understand them (the masses), so that they do not even know that the complexity exists. Perspectives fall under this category. The UI should unfold for those for whom the basic functionality does not suffice.

            I fail to see how one caters to the masses without first catering to the people who will feed the masses. If 1% of the users create the content that 99% will consume, what happens if the 1% leaves? It's not about choosing to please advanced or basic users, it is about empowering the builders to build so the rest will have something to come to see. It's very easy to build a tent, and pretty straightforward to use. I'd rather build pulleys and levers, forges and forms and let the architects build palaces. But hey, tent cities are great, aren't they? It's just a matter of where to take a s&*# that's complicated, oh wait, that should be simple right? If the architects are given the right building blocks, they will build the drainage systems.

            In pursuing simplicity, one can either stop at doing simple things simply (and be crushed by the large group of competitors in this space) or dare to push through complexity to get to a new simplicity. I hope twine will do the latter.
            Aasemoon'z Twine

            • Nova Spivack Nova Spivack
              9 months ago


              I agree on your first paragraph totally. That is where the focus is right now.

              For your second paragraph, I would love that, and we are not too far from some of what you want. Everyone would benefit from this as well, but first things first.

              We're in agreement on this.
              Aasemoon'z Twine

          • 9 months ago


            Kurt and I seem to have slightly different priorities on our desires for Twine, but in the end they are approximately the same. My primary use of Twine relates to semantic content creation. I want to create rich, meaningful content that I can use in other applications outside of Twine, as well as within Twine. Services such as Dbpedia and Freebase, while excellent, are geared toward a certain purpose whereas Twine offers a much more open environment for content creation. I also hope to see the ability to mix ontologies or possibly even create or include new ontologies for additional content options. While I enjoy the promise of machine-added semantics, I am happy to create them myself and would like to see that option restored.

            My second priority would be to see Twine items thread Twines together. I'm generally less interested in which Twine I'm currently viewing and think that a "Twine" would be better used as a tag or categorization of items, i.e. just another ontology for delivering interest-related items to users. I think this aligns with Kurt's second "level" and possibly a bit with his first. Perhaps one might say I view a "Twine" as a superset of related "tags", but that wouldn't be quite right either. I see a great benefit to the grouping of items in a Twine outside of tags; however, I think they are merely an ontological grouping assigned by users, with a potential for machine-generated sharing.

            Finally, I absolutely agree with Kurt's final two paragraphs. The "masses" will most likely try to use Twine as a combination of StumbleUpon and Delicious. The very great probability is that they will find those tools more satisfying than Twine and go back to them (as I did with Delicious for a time). The one thing that might hold them here is the community, but they first have to get involved. A great many content items are just that--no comments or discussion, just bookmarks, etc. Also, given the feel from my other reading surrounding the SW, the majority of the people who will really love Twine are the Power Users. I know that cuts into potential users, revenues, etc., so if that is what will keep Twine going, then I think all of our idealism will be let down. Realism, given such constraints, means that much of what we hope for will continue to be pushed back to a later date, ad infinitum. The balance is indeed tough to draw out; I just hope that we'll see a swing towards the builder-types so that Twine's future will be secure as a beacon of semantic goodness, to wax poetic... in a way. :)
            Aasemoon'z Twine

            • Nova Spivack Nova Spivack
              9 months ago


              Ryan the good news is that what you are asking for in paragraph 1 is something we literally are working on -- but it may end up becoming a separate OSS project. It's a "free time" project right now so it might take a while. In any case we do have something that literally does everything you ask for. The question is -- what are the use-cases? Who would use it and for what? Assume you would be able to mix any ontologies, or make your own, and post any structured content to any site, with good RDFa, and everything else you would expect. And then search it. Is this something that would be widely needed? In any case, we have it in alpha right now, and hope to release it this year somehow.

              As for your second paragraph -- would love to hear more suggestions about that topic. Twines basically provide focused group, search, and moderation functions on some collection of content. They are really wrappers that combine collections and people. I think they should be linkable, combinable, stackable, etc. Also there ought to be container objects within them too. But these are definitely less urgent features than, for example, making "My Items" more useful.

              As for paragraph 3, maybe or maybe not. For simple bookmarking, perhaps. But "bookmarking" is not where Twine is ultimately headed. We're really going towards "interest tracking" -- making it possible for people to keep up with and discover content around their interests. I'm not sure Delicious or StumbleUpon really work well for that purpose.
              Aasemoon'z Twine

          • Bent Rasmussen Bent Rasmussen
            9 months ago


            Features that were removed that were of great value.

            Well....

            - I really liked being able to have a series of images within one image item. In general multiple properties of the same type on the same item. I know this was a major one for Aasemoon as well

            - I really liked being able to bookmark twines into other twines; this was actually a key use-case that was tossed out - apparently by some rationale; but it is a key use-case that makes, even if it didn't quite have the UI/UX around it to make it very compelling.

            Those were two immediate things that came to mind but I wrote a "rigor mortis" note in the lounge about it.... Features that were removed that we really appreciated.

            I don't have anything specifically wrt search except that being able to have domain-specific search would be compelling; searching for book prices, movie directors etc. (well there's that idea of IMDb template support)
            Aasemoon'z Twine

            • Nova Spivack Nova Spivack
              9 months ago


              Multiple properties of the same type on an item could probably be easily re-enabled. As you may recall the GUI was kind of klunky for that and since few users were using it we opted to remove. We'll bring it up in a product meeting and see what the level of effort would be to add back in. It would be a nice way to have more semantics in the surface level app. Is this important to more people?

              Ability to bookmark a twine into another twine worked last time I tried it. Didn't know it went away. Anyway, I agree that is useful and should be easy to re-enable if other people want that.

              I'm not a member of the Lounge so I can't see what is being posted there -- but maybe it is time for me to be invited in. Hmm.

              I think search on semantic properties would be great and we've discussed that too. To do it well is a big effort. Most users don't use advanced search on any search engine. So we've opted to focus first on things like My Items, Item Detail pages, etc. which everyone uses. Then turn to things like this. But again, a great opportunity to show more semantics here.
              Aasemoon'z Twine

              • Kurt Laitner Kurt Laitner
                9 months ago


                I will say it once more, then leave the point, volume of interest alone does not make or break a feature, compelling leverage is far more important - maybe 10 people will understand a feature and its implications enough to use it, but if their use of this feature provides simple and compelling content for other users (see level three above in my levels of UX comment) it is worth doing. Florian and Bent would be creating fantastic things if they had properties and search by properties combined with feeding search into a twine and twining a twine having he additional feed behavior option. I might even have some fun with it. And the key is, the resulting structure (a twine) is a consumable. You've got a small dev team, you should be doing everything you can to make it bigger by co-opting the enthusiastic participation of your core users to prototype for you.
                Aasemoon'z Twine

              • James Todd James Todd
                9 months ago


                we're gearing up to take on bookmark a twine resource but do it the right way this time which will in effect be equivalent to a on twine share. eg bookmarking a user, item, spot can easily map to an on twine share ... noting that only the item share is supported in the ui today.

                i suspect the bookmarklet front end will be agnostic as to the backend details.

                not sure about bookmarking non-object pages though (eg /search). thoughts?
                Aasemoon'z Twine

                • Kurt Laitner Kurt Laitner
                  9 months ago


                  re 'bookmarking' search results, an optional behavior needs to modulate that action (when you bookmark the search results, are you indicating to share all items in that resultset to the twine noted, stream further results from that resultset to the twine indicated, or provide a link to the search result page in the twine indicated...)

                  looking forward to playing with this one before commenting further.
                  Aasemoon'z Twine

                  • James Todd James Todd
                    9 months ago


                    yup, that one is out of scope ... just to be able to turn the corner on this pup.

                    feedback, though, especially up front, is totally welcomed.

                    thinking about bookmarking collections (specifically search results) is problematic on a number of fronts. will keep pondering.

                    lastly, i personally was hoping the "user forum" twine would be the place to have discussions like this. i had liked the "twine feedback forum," initially, but the signal-to-noise got, candidly, horrendous. we want/need/request/desire/beg/love feedback :)
                    up front when at all possible.

                    should we start a new "user forum" thread(s) for details like the one we are discussing.
                    Aasemoon'z Twine

                    • . .•∞º˙ . .•∞º˙
                      9 months ago


                      WHAT??? And ruin all the fun?

                      THIS kind of thread is why I got hooked on Twine in the first place! I don't know a lick about semantic engines and RDF, or ATOM feeds, but I do know that exchanges like these are what make Twine. Don't jump to another thread. Build on this one and make it sticky. You start a magnet feed like this and eventually all sorts of people get involved who would never even have taken interest before. That's why I really enjoyed TBFAS/TUFAS--I learned a heck of a lot about what you guys are trying to do, and made some great connections while participating there. (I mourn it's passing).

                      Don't break away from this thread and start a new one. Bring us moderated controls. Get us Reputation-based comment visibility. Make it so big comment chains like this can be rolled up in different ways beyond time-value or comment-locality. Show me last comments first. Give me blocking capability to not have to read undesired member comments. Set comment views to only show unread items instead of all items every time. Set up Thread-level controls to filter and review a long chain like this in all sorts of possible ways. Heck, I can do that with Windows Folders now, and that's web 0.1...

                      Make the system easy to use and collect the valuable nuggets any way you want, and still have all the silliness. THAT's how you grow eyeball counts.
                      Aasemoon'z Twine

                      • Kurt Laitner Kurt Laitner
                        9 months ago


                        you forgot place me back where I was in the thread, but well done nonetheless - oh crap, as soon as I hit reply I'm going to be transported 1000 miles away.. away away remember me
                        Aasemoon'z Twine

                      • Kurt Laitner Kurt Laitner
                        9 months ago


                        further to rick's comment - and the friendship in hyperconnectivity and polylogue discussions there, threads of any significant interest and contribution level get messy, and wouldn't it be fabulous to be able to deal with them - this would make twine unique in my mind, a competitive differentiator, and quite frankly a service to sell if properly modularized. a way to clip a branch at any node and start a new discussion (either by the moderator only or by any participant) as well as a tunnel between the two resulting views that allows either hard forking (both continue independently) or soft forking (the second view is actually just a filtered view of the first for those only interested in that sub thread) would be fantastic.

                        And yes, this thread is old school Twine...
                        nooooo school like the ooooold school.
                        Aasemoon'z Twine

                        • . .•∞º˙ . .•∞º˙
                          9 months ago


                          Ahhhh....old school....I remember those lovely Twines, 4 score and seven months ago.

                          And this one's up to 121 comments (122 with mine!) WOOOHOOOO! Ridem cowboy!
                          Aasemoon'z Twine

                      • James Todd James Todd
                        9 months ago


                        in principal, i agree and that is eventually possible.

                        it is just that there is no logical reason a child item/note can't be created in a more contextual area to which the details can be layed out w/ little to no room for misinterpretation. user forum is the spot for this type of dialog at this point in the journey, imho.

                        bigger picture, what you describe is entirely possible (read threads of interleaved twines) and dearly needed (i would love to be able to filter comments by user, or, better yet, have it filtered for me based on reputation/etc) ... it's just that at this point and time and on this specific topic (ie bookmarking twine) i believe it merits it's own top-level talking point that can/should/may be referenced elsewhere. there are many a talking points that can be given the same treatment. heck, i'll probably try to start making it a practice to seed such a discussion in user forum moving forward and simply reference it when needed.

                        hth!
                        Aasemoon'z Twine

                    • Kurt Laitner Kurt Laitner
                      9 months ago


                      oh yeah, and @James, you've had that feedback for almost a year now (see Beat That Horse and sub twines for many treatments of that issue from multiple angles - good bed time reading I'm sure), speaking of which I need to post this thread there
                      Aasemoon'z Twine

                    • Kurt Laitner Kurt Laitner
                      9 months ago


                      further to rick's comment - and the friendship in hyperconnectivity and polylogue discussions there, threads of any significant interest and contribution level get messy, and wouldn't it be fabulous to be able to deal with them - this would make twine unique in my mind, a competitive differentiator, and quite frankly a service to sell if properly modularized. a way to clip a branch at any node and start a new discussion (either by the moderator only or by any participant) as well as a tunnel between the two resulting views that allows either hard forking (both continue independently) or soft forking (the second view is actually just a filtered view of the first for those only interested in that sub thread) would be fantastic.

                      And yes, this thread is old school Twine...
                      nooooo school like the ooooold school.
                      Aasemoon'z Twine

              • James Todd James Todd
                9 months ago


                the multiple images per "image object" page largely boiled down to legacy issues on the (way old) item detail page, which is gearing up for a revamp. i believe there is a difference between "an image" object and "an album of images" object. perhaps end users need only see the later? i know that is what my wife expects when using twine and sites with like functionality.
                Aasemoon'z Twine

                • Bent Rasmussen Bent Rasmussen
                  9 months ago


                  There is a difference, correct.

                  The problem is that there is really not many templates of expression. Realize that as a user I am not caring about the formal hygiene of the data. This must be taken care of transparently.

                  What I care about are the right presentation templates that allow me to customize the end-user experience of an item.

                  Transparently, any underlying details about images etc. is a separate layer. Maybe you're coupling these concepts too tightly.

                  A UX concept should really be a separate composite model of some set of entities arranged visually.

                  Ideally we'd be able to

                  1 - create our own visual models
                  2 - create our own logical models
                  3 - bind the visual models to the logical models

                  That said, just some awareness around the end-user UX of item presentation would be a leap forward. Because the item/list metaphor is a bit stale.

                  Just a few thoughts...
                  Aasemoon'z Twine

                  • James Todd James Todd
                    9 months ago


                    concur.

                    the system is largely data driven, which also makes it rather sterile/generic (imo). each of the pages are not getting alot of focussed treatment which will challenge this approach, for good reason, as all pages should be approachable and something one can personalize (which can have a number of meanings).

                    seeing the changes ripple throughout the site, i'm very optimistic the item-detail page (that we are on now) will be much improved in the coming weeks.

                    thoughts welcomed/appreciated.
                    Aasemoon'z Twine

                  • Kurt Laitner Kurt Laitner
                    9 months ago


                    "1 - create our own visual models
                    2 - create our own logical models
                    3 - bind the visual models to the logical models"

                    yes!!, see also graph oriented programming and node-component pairing concepts for more on this - effectively a composite node 'type' can have a component of programming hung off of it, containing the behavior and representation of that node - lego for structurally minded non - programmers or lazy programmers.
                    Aasemoon'z Twine

                  • 9 months ago


                    Reminds me of Drupal's custom content type (or whatever it's called; I haven't used Drupal in some time). I think WordPress even has a similar plugin now (http://wordpress.org/extend/plugins/custom-field-template/). Those might be helpful in looking at simple ways to include this kind of functionality. Granted, they'd only be a start.
                    Aasemoon'z Twine

          • François Dongier
            9 months ago


            I think the core of my frustration (why I feel things are moving too slowly) has to do with the poor quality of the automatic tagging. The entity-extraction algorithm used in Twine just doesn't seem to be the best on earth. I think this is important because most of the interesting semantic features we're dreaming of (like good connections recommendations, or good reading recommendations) cannot be realized until this algorithm improves. People, places and organisations are extracted pretty much correctly. But the "Other tags" (which ideally should extract concepts and ideas the article is about) are, IMHO, still very very bad. And these "other tags" are of course the most important ones.
            OpenCalais does much better. Try any page of your choice on http://www.semanticproxy.com/demo.html and then on Twine and you should see the difference. I haven't seen Primal Fusion yet, but I believe it uses Yahoo! BOSS and so should also do a lot better than Twine.

            Just a couple examples, from a couple of twines I manage (Microfinance and Primal Fusion).

            Other tags for Primal Fusion = {arts, business, canada, demo, food, kitchener, news, newspaper, ontario, primal fusion, record, search, semantic web, semantics, sports, the}
            Recommended twines: {Local Food, Nutraceutical, Chefscorner}.
            Hilarious!

            Things are better in the Microfinance twine, but that's because some contributors are very conscientious manual taggers.
            Recommended twines = {Economy, Sustainable Development, Marketing}
            Other tags = {africa, bangladesh, business, finance, grameen, innovation, leadership, loans, micro lending, micro loans, microcredit, microfinance, microloans, nobel peace prize, poverty, web 2.0}
            Because of the quality of the manual tagging, reading recommendations are much more relevant.

            This being said, I am confident that this problem can easily be be improved. And when it does get improved, Twine will become really interesting. So, to summarize: yes I am frustrated, but still very optimistic.
            Aasemoon'z Twine

      • Bent Rasmussen Bent Rasmussen
        9 months ago


        Let me chime in, as I drew the ugly picture.

        I get the release cycles are short but still, in the total picture it just looks to me like several key things were removed or broken while some things were of course added. This leads me to see a picture of stagnation. And this is where all my alarm bells and whistles finally came out of the closet. They came out in an explosion and I hate this - it was only because of the significant personal investment by me and others that I wrote this (and even external encouragement), otherwise I would have silently abandoned the ship. Such a customer, I'm told, is the most dangerous customer - the customer that doesn't let you know what you're doing wrong, but just never comes back after being disappointed once.

        I will say that my criticism was quite harsh and I've probably offended several people, although it was not targeted at any specific individual, more on the situation as a whole. (Including the lounge, which is a good idea in theory and I have nothing against Bonnie Ho whatsoever; but being a member there with whatever you say you can only get the response that the team is looking into it, it's not very interesting to be honest; maybe I was expecting too much and it would have been better to just not be there in the first place; which was basically my conclusion.)

        As for the API issue. Well, personally it isn't so much about advanced querying capabilities as an open REST based API. And basically just completing the RDF exposure across the site. The RDF support is spotty and full of holes. Some places there's RSS or Atom support, some places there's RDF support. Some places neither.

        But to be honest, seeing how things are I can't help but feel desperate about these priorities (yes, I'm in no position to ask anything, I am just stating how one so called power user feels on the matter) -

        1 - that bookmarking speed be fixed and come back to normal (the same goes to the on-site item forms which are excessively slow)
        2 - that the explore new twines pages be put back with a moderated queue system, such that twines are manually checked before they can be seen in the view (by staff and/or community members); for the life of me, I cannot fathom the challenge in that - simply monitoring new twines and putting a checkbox on the ones that are green; sure maybe they need to be there for a while, with 5-10 genuine items before they become visible, to completely elliminate the possibility of spam; I know - more work

        I can't access my interest feed and haven't been able to for days. Deleting cookies or changing browser does not help! But of course, I'm one of those rare cases with more than ten thousand items. Or what is it, I can't find the statistics for it anymore...

        Features trump speed but if speed is this dismal, it becomes the overarching concern that detracts focus from everything else. If new features are checked in that degenerates the UX so severely maybe you should reevaluate the benefits of these features. Or maybe just do some more testing and use performance metrics before landing them. I don't know; wohn't try to play know-all here, I'm just a biased observer with a lot at stake.

        I think having to wait 3 weeks (or maybe several sprints?) for a critical performance fix is not what you want. Trust me, it's not in your own best self-interests. I can't believe I'm alone on this point. I know many people who would not even join Twine with it's initial speed. So now it would be like asking a regular at the Waldorf Astoria to relocate to a soup kitchen for the homeless (no arrogance intended). Again, speed is not of the essence, but there is an essential speed floor, under which working with Twine becomes a showstopper.

        The areas of future improvement you mention sounds encouraging Nova. I'll be a passive observer from now on. Lettings things speak for themselves. Experience is the best teacher anyway.
        Aasemoon'z Twine

        • Nova Spivack Nova Spivack
          9 months ago


          Thanks for the feedback. All good suggestions. I agree bringing the New Twines page back, perhaps as you suggest, by only allowing twines that have achieved some level of activity and endorsement (from other high ranked users, or at least more than just the twine creator) would solve the moderation issue there. We will talk it over. As for the API and RDF: yes. Agreed.

          It sounds like you are having two issues that are related to the size of your account, and possibly some kind of bug as well. Bookmarking speed has been buggy in the last release. But I hear it has been substantially improved in the next release (later this week). We will see. For users with less than thousands of items it will probably be fast in all cases -- only those of us with enormous accounts will see these issues generally because we push the platform so hard.

          That said, it is important (to me too!) that these issues get resolved so even those of us with enormous numbers of items in Twine can get great performance. The team is very aware of this. The solution is to tune and upgrade the data model and database, which affects the application at many levels and touches a lot of code and hardware. That does take many sprints to plan, implement, test and release, but keep the faith, it is improving steadily and will get much better in the next few sprints.

          As for you not being able to access your interest feed, that sounds like a bug and I wish I had heard about this sooner (I just found out about your issue from Kurt yesterday). I will expedite a solution to this. Very sorry to hear about this.

          Meanwhile, try this experiment: create a new test account and see what it is like for a new user. You will notice that speed and so forth is much better -- certainly acceptable at least. The slowness you have noted seems to kick in when accounts get very very large. This is something that the architecture improvements will fix.
          Aasemoon'z Twine

        • James Todd James Todd
          9 months ago


          i can reproduce your inability to access your home/interest page. looks to be a data issue related to your non-default setting of displaying items by time, vs by twine. we need to capture a data dump that in can inspect internally in order to resolve this one, i suspect. hmmm, i wonder if we could reset your view to the default (group by twine) if that might get you going in the mean time. would you be opposed to that?

          re atom, every object and collection of objects in twine has an atom representation. we don't yet publish atom service documents (that make it easy to see all the feed sources) or atom pub (read/write) at this time. something to consider moving forward though, as in how interesting/important would each be?
          Aasemoon'z Twine

          • Bent Rasmussen Bent Rasmussen
            9 months ago


            Feel free to go ahead James.

            More of an RDF man myself. RDF exposure is more meaningful (obviously).
            Aasemoon'z Twine

            • James Todd James Todd
              9 months ago


              sweet! we'll give this a go shortly. i still want to get a data dump in house and sort out which comment object is running amuck. but that need shouldn't block your use of twine in the mean time.

              i hear ya re rdf. personally, i believe an extended atom doc w/ rdf overlays is serious mojo. add in atom service docs/atom pub ... well, that would be a fun data source to surf through :)

              now, saying the above does not imply that rdf will necessarily be imprisoned w/in atom alone ... it's just that the 2 technologies combined provide powerful synergies re data lifecycle management.

              we'll get there.
              Aasemoon'z Twine

              • 9 months ago


                +1 RDF in Atom would be a dream come true. Well, even exposed RDF would be nice. I've been working on a small Silverlight Atom publisher (probably one of many) that I would love to use with Twine.
                Aasemoon'z Twine

                • James Todd James Todd
                  9 months ago


                  we're starting to build on some of these ideas (in a upcoming rsn feature) which i think ROCKS!

                  we've run into some kinks (i love working the unknown though) ... but we're well beyond crawling in this regard.

                  parallel, though, i think exposing an atom/rdf data set will be pretty trivial and a foundation that can be readily built upon by a wide set of tools.
                  Aasemoon'z Twine

            • James Todd James Todd
              9 months ago


              are you back in now?
              Aasemoon'z Twine

        • Scott White
          9 months ago


          All feedback is good feedback. Although constructive feedback is preferred, sometimes too much negative feedback that could have been communicated constructively is a sign that there is a listening process that is not working. It is my hope in coming weeks Twine can improve its ability to listen in a way that users feel they are being heard. Hopefully, this will disincentivize users from feeling that throwing a bomb over the fence is the only way to be heard.

          Won't comment on the other points as Nova and James have already addressed them.
          Aasemoon'z Twine

          • Kurt Laitner Kurt Laitner
            9 months ago


            "sometimes negative feedback that could have been communicated constructively is a sign that there is a listening process that is not working"

            astute insight, hope you don't mind if I borrow that for future use, nicely put.
            Aasemoon'z Twine

    • Greg Boutin Greg Boutin
      9 months ago


      I know Nova's smooth reply wasn't directed at me for sure. "We have no problem with people giving us constructive feedback". Yeah, right. Problem is, criticism based on perfectly valid data provided by MC Nova himself is not considered "constructive feedback". As far as Twine goes, constructive feedback is positive feedback. Anything short of that, and Nova hangs you publicly on Twitter with personal attacks accusing you of dishonesty and "hidden agenda" in a totally paranoid fashion (check out the thread for yourself).

      I know he's going to argue that I misrepresented his data... bla bla. I reproduced all the data he sent me as he sent it to me, and then chose in my introduction to focus on one particular metric and express my opinion based on the january's number that I had at the time of writing my post (he provided the other data later, and I didn't update my number from 2 to 3.4. Woo ooo, big deal - note it was down from 16 views/visitor... so on that scale it's really insignificant. Just an excuse for Nova to get rid of a person-non-grata).

      I think fishhead's analogy of China with Tibet is right on. All signs point to Nova only caring about how it's going to look with his investors so he can cash out (before the ship sinks?), and that's why he's trying to keep you as quiet as possible. Try going public with your comments, and see what happens for yourself... Your "constructive" feedback will be re-labeled in no time...

      In the past, I've worked for years writing published market studies as a diplomat, got my mba from stanford, and worked at the boston consulting group, and I only mention that to say I never once have been criticized for not getting my facts straight. Everyone can check for themselves by seeing the twine post on my blog.

      Sorry, I don't buy it. Burnt once...

      Frankly, I even suspect this answer wasn't actually written by Nova (unless he's totally bipolar). Hi Candice.
      Aasemoon'z Twine

      • Nova Spivack Nova Spivack
        9 months ago


        Greg, thanks for your comments. But the fact is I and others did a lot of work gathering and giving you data and responses you requested. It was a disappointment to then not have you quote the data we gave you after all that. Had you simply replied that you would update it when I commented, instead of saying, "what's the big deal?" I probably would have been less disappointed.
        Aasemoon'z Twine

        • Greg Boutin Greg Boutin
          9 months ago


          Nova, that "disappointment" you keep quoting led you to totally unacceptable behavior with a number of personal attacks. You're not worth of your position. Your expectations were yours. I had told you I was going to publish the interview entirely on my blog, which I did. The rest is my own analysis, I presented it as such, and everyone knows that. Following the venturebeat article, I gave you a chance to express your view, and quite stupidly you decided to treat me like a nuisance to be dealt with, just like you had done when I and a number of other users expressed "constructive feedback". In front of my twitter followers, you have put my independence in doubt and questioned my integrity, accusing me of "hidden agenda" without a beginning of evidence. I think the history of my blog post proves my independence, and your history of controversy on Twine and undemocratic reactions hints to your lack of objectivity.

          On "hard facts", the fact is, you're brushing off that average # of pageviews as irrelevant, and that's your right, and it's mine to believe and express that the CEO of a web company should be a little less dilettante about such trends. Your strategy is your own, but whatever the obscure "power laws" "dictate" (sounds almost Asimovian), I have dealt with enough websites (recall, I was marketing director at a web venture, and dealt with the web for over 10 years) to believe this is a significant number that echoes my experience-based impression about the usability issues of your service. Registered users is one thing, and even it has gone down. Fact is, other folks come to your site and choose not to register. Doesn't that say something about your retention? Will that not have an impact on your revenue, when you put your product recommendations on...? So please, if really you believe that's irrelevant, leave the numbers to someone who can handle them...
          Aasemoon'z Twine

          • Nova Spivack Nova Spivack
            9 months ago


            Greg, you just tweeted that, you "would rather not be seen as a Twine power user" on Twitter -- So that pretty much speaks for itself to everyone here.

            And, once again, your facts are wrong: You state above that "registered users has gone down but actually, again, you are incorrect, registered users has gone significantly up. You can go ahead and call me pointing that out a "cheap shot" but I don't think anyone here is going to buy that. If anything I am trying to be objective and I'm asking you to do the same.
            Aasemoon'z Twine

            • Greg Boutin Greg Boutin
              9 months ago


              mmm, you'll excuse my foreigner's grammar, but #pageviews for registered users has gone down is what my sentence is intended to mean, which is obvious in the context of our discussion as you know, since it's in my post and on your tweets.

              As for "I would rather not be seen as a Twine power user", you also know it came as tongue-in-cheek response to you telling me I was about to be removed from power users due to my "cross-posting", which btw is something I was told by your own people is not a problem (I inquired b/c I was worried this might be seen as spam, but was told as long as I cross-post on relevant twines it's ok).

              The cheap shot I'm accusing you of is about your unprompted tweet saying "Nobody's forcing you to cross-post links to your blog article 14 times on Twine, Greg."

              Your distorted view about "average not making sense" is only a self-serving perspective, and again you're free to explain as much as you want where your focus is, and I'm free to say your retention is terrible.

              So stop rearranging comments to serve your evil purpose. You're so totally corrupted by your trade I don't expect one inch of intellectual honesty from you. I will leave it at that until my next post on Twine. Feel free to torpedo as much as you want here in the meantime.
              Aasemoon'z Twine

              • Nova Spivack Nova Spivack
                9 months ago


                Never said you were going to be singled out. I said you found a flaw in the algorithm. You were correct, you should not be included given that you really don't use Twine much. From what we can tell all you have done is cross post the same item many times recently. The algorithm should not count those as separate posts. Nothing wrong with cross-posting to relevant twines with relevant content, within reason of course. But the algorithm should not be ranking you as an active user for that. Thanks for pointing out that flaw in our algorithm.
                Aasemoon'z Twine

    • Greg Boutin Greg Boutin
      9 months ago


      I almost forgot... For those among you looking for alternatives to Twine, I suggest http://www.noovo.com/

      For full disclosure I have zero interest in Noovo, be it financial or any other. Just interested in providing new options to disgruntled Twine users like me.

      I haven't been using Noovo much so far, but I'm sure going to give it a second look.
      Aasemoon'z Twine

      • Kurt Laitner Kurt Laitner
        9 months ago


        Hmm. a list of interests on the front page...

        From top left..hot on noovo right now:

        Smart babes profile
        Myspace, How They Started
        People are Slinkies
        Why Handsome Men Make Bad Husbands
        Wealthy Men Give Women More Orgasms

        This is where pandering to the unwashed gets you Nova, but hey, maybe they're making more money.

        Good news though, I'm a great husband.
        Aasemoon'z Twine

    • Greg Boutin Greg Boutin
      9 months ago


      ok everyone on twitter now please, nova is at it again. @novaspivack and @gregboutin
      Anyone care to help bring some objectivity?
      Aasemoon'z Twine

      • Nova Spivack Nova Spivack
        9 months ago


        Dude, what are you talking about? You posted that you found it odd that you are ranked in the top 100 users because you almost never use Twine, as you yourself say. We looked into it and found that you are right, you really don't use Twine much (which of course begs the question of why you care so much about Twine in that case, but anyway...). You are only ranked in the Top 100 Users because you cross-posted links to your blog article to 14 twines at once, causing our ranking algorithm to think you were very active. Thanks for pointing out that flaw in the algorithm. It will be fixed shortly.
        Aasemoon'z Twine

        • Greg Boutin Greg Boutin
          9 months ago


          the cheap shot I talked about. Can you please also single out other folks who cross-posted their posts? Please mention the number, like 11, or 12, or 14. What a lonely man you must be.
          Aasemoon'z Twine

          • James Todd James Todd
            9 months ago


            i'm tiring, excessively so, of the trite and disgusting personal attacks greg. pathetic.

            can't wait for reputation based results whereby noise is simply and organically buried.
            Aasemoon'z Twine

            • Greg Boutin Greg Boutin
              9 months ago


              wormtongue... you accused me of seeing conspiracy and then can't come up with one fact to back it up. All you could do was come up with another cheap shot on technical lingo. So please, don't come preach about personal attacks here.
              Aasemoon'z Twine

              • James Todd James Todd
                9 months ago


                oh ... nice. lotr reference? i see myself more as aragorn :)

                so buried in "your" facts you readily loose context, or choose too.

                you: twittered "excessively" about not being able to log into twine and quickly came to the conclusion you were "blacklisted" ... for the second time.
                others: told you to dial down the rambling paranoia, as they, too, couldn't log in
                twine: site was down and will be back up shortly
                me: stop looking for conspiracies and start looking for the truth

                techno lingo? interesting. preaching? please! personal attacks? get real!
                Aasemoon'z Twine

            • . .•∞º˙ . .•∞º˙
              9 months ago


              "...reputation based results..." Is that something in the pipeline??? WHEEEEEEEE!!! Can't wait for that one!
              Aasemoon'z Twine

          • Nova Spivack Nova Spivack
            9 months ago


            When the algorithm is fixed everyone will be treated the same way, not just you. It will discount cross-posting for everyone equally. The algorithm should weight unique postings more than cross-postings. This is a flaw, and we appreciate you helping us to find it and fix it.
            Aasemoon'z Twine

          • 9 months ago


            I have to agree with Nova here, Greg. I don't see a cheap shot in the mix, but I do see two people who are quite agitated with one another. From both perspectives (as best as I can place myself) you see yourselves in the right with excellent points. However, trying to highlight every little thing Nova says as evidence of your case (when objectively it isn't) is a little juvenile. I (purposely) avoid Twine, so I can't comment on that, but given the three posts above, I'd have to say you need to ease up a bit. You appear to be picking a fight, even while claiming that the fight is being brought to you.
            Aasemoon'z Twine

            • Greg Boutin Greg Boutin
              9 months ago


              ok thx for the feedback, Ryan, and at this stage it's true i am really agitated by this. There has been too many personal attacks and sneaky attempts at going after me based on other things than facts. I'll ease back but I would appreciate it if you could observe the number of unrelated attacks I have been the subject of after publishing the interview, and realize Nova should not have taken that to Twitter with such attacks.
              Aasemoon'z Twine

    • Greg Boutin Greg Boutin
      9 months ago


      Good for you, Nova. Now people please go see the thread please and make your own mind.
      Aasemoon'z Twine

      • Nova Spivack Nova Spivack
        9 months ago


        Thanks Greg, I'm sure they will. And anyone who is curious about digging deeper into your run-ins with other CEO's and Semantic Web thinkers can just look at this thread, which speaks for itself: http://www.web3beat.com/2009/02/the-definitional-challenges-of.html#comment-8377 The folks you are fighting with there are some of the biggest names in the semantic space.
        Aasemoon'z Twine

    • Greg Boutin Greg Boutin
      9 months ago


      yet another cheap shot unrelated to the fact I exposed. yes thanks for referring that thread, everyone feel free to take a look. We are debating the semantic web and linked data, and the attempt at reframing are plenty. Note that came up a day after nova's outburst, and the so-called "semantic leaders" are clearly referring twine's debate... Tribe uniting together... Stop the intimidation policy.

      I repeat: I published the interview entirely on my blog. The rest is my own analysis, I presented it as such, and everyone knows that. Why do you feel so threatened Nova, if as you say my reputation is so low now?
      Aasemoon'z Twine

    • Greg Boutin Greg Boutin
      9 months ago


      ok i'm leaving this thread as I wanted to do yesterday. Apologies to readers for all the bad feelings, but really the Twine approach to 'constructive feedback' needed to be illustrated a bit further.
      Aasemoon'z Twine

    • Bent Rasmussen Bent Rasmussen
      9 months ago


      Nova, did you try to post a comment to my blog? I can assure you no comments were removed by me, although they often go into a queue to be approved. I have no pending items as of yet, so maybe a Wordpress glitch. Too much text to read at work, digesting later...
      Aasemoon'z Twine

      • Nova Spivack Nova Spivack
        9 months ago


        Yeah I did respond with the same big answer above, so this is fine if you just want to link to this thread. I also submitted just the link to this thread for you. Neither got through so maybe there is a comments issue on that post? Thanks for writing. Hopefully this all helps a bit. Well at least the part of this thread that is about actual power users with actual feedback :)
        Aasemoon'z Twine

    • Aasemoon . Aasemoon .
      9 months ago


      Wow... oook! So I didn't check this thread for less than a day and it seems to have grown a few meters.

      First of all it's definitely great to finally hear from Nova. Reading about the plans at hand makes me feel a bit more optimistic now. More of this, would CERTAINLY be appreciated! It's one thing to see features disappearing and the tool becoming less and less reliable and more and more difficult to use every week, but it's a whole different story to feel completely blind of the bigger picture. Not knowing whether to abandon ship or wait some more and keep hopes up that the cracks will be sealed soon enough. Now at least I know that there's still hope...

      And I don't quite know what to think or say about the "discussions" here... as all parties seem to have valid points. I feel that at this point only time can clarify the validity of certain things.....So since I'm feeling more hopeful now as I said, I guess I'm going to stick around as usual, and see what happens... in time. Hopefully in my lifetime.
      Aasemoon'z Twine

      • Nova Spivack Nova Spivack
        9 months ago


        Thanks for the response, Aesemoon. Your contributions to Twine are extremely valuable, and so is your feedback and opinion. We are going to be meeting at the office about how we can keep you guys more in the loop, and gather feedback from you, more effectively. We're also looking into Bent's account problems, which I was not aware of, and I hope we'll have a solution for that soon.
        Aasemoon'z Twine

    • Bonnie Ho Bonnie Ho
      9 months ago


      An Update on the System Issues:

      Just wanted to let everyone know that we've identified a network configuration error at our provider that is causing the frayed messages. We expect it to be resolved within 24 hours. We will keep you posted if anything else comes up - but hopefully, you'll just see the results. :)

      Thanks for your continued support!
      Aasemoon'z Twine

    • Nova Spivack Nova Spivack
      9 months ago


      Just wanted to update everyone on when the new features (new interest feed alpha, etc.) that I referred to above are coming out. We were going to release them this week, but given the other technical issues that came up this week, the team release has been delayed. This is good though - -it gives the team the rest of this week to do a little more testing first. So these features will release early next week instead.
      Aasemoon'z Twine

      • 9 months ago


        Thanks for the update, Nova. Just curious, do you have the ability to pre-release certain features only to certain members for testing purposes? I realize the whole site can't continue in beta forever, but I would enjoy helping test some of the features I'm (we're) requesting here, even if it means a bit of pain in working out the details. I generally don't mind pain, at least not when it concerns either sword fighting or testing. :)
        Aasemoon'z Twine
    Add a Comment
Report This

Twine is about discovering, collecting and sharing the content that interests you. Learn More

Join Twine

Stats

First Posted By

Bent Rasmussen

First Comment By

Aasemoon .

Tags

Community Tags

Who's Interested In This?

Forgot your password?