Please evaluate the new forum software (with a test account)

Now that we have successfully migrated the wiki to the new server, it's time to finish the job and get a forum running on the new server.

For the benefit of newcomers, here is the background story. Until this time, Andrew Stacey has been maintaining the forums and the wikis for the Azimuth Project, and the nLab. For Azimuth, we are now in the process of transitioning those responsibilities to me. For the wikis, he had a VPS which supported both the nLab and Azimuth, but that VPS is now being run by someone else for the nLab, and we now have our own wiki server on a hosted account. The Forum is currently running on a server at the university he is at, but he has relocated, and we no longer have a guarantee of ongoing support on this server. For today, we're okay, but we need to keep moving, in order to make sure that we stand on our own feet, using servers that are under our control.

Also, Andrew has limited availability for working on this stuff now. Fortunately the forum is running on its own really well now. But that does mean that it is running on autopilot, and I am not in a position to administer it since it's not running on our server.

Brace yourself, folks, but it looks we're going to let go of this forum software, which has served us quite well, and move to a new package, which also looks promising and which offers us the prospect of ongoing support from the developers. We'll have to be prepared for some change, but I believe it will provide us with substantially similar functionality.

The current Wiki starts with Vanilla 1.1.10 (a open-source php forum package), and contains further customizations that Andrew wrote. Now the current version of Vanilla is 2.1.1, and, to get current, we would have to work to port his changes -- and there may be compatibility issues between his changes and the new code. The alternative, of staying frozen at 1.1.10 does not feel right to me, because software does tend to get rusty as the things around it change. Moreover, I don't know anything about Vanilla, or the customizations, and don't want to take on the job of maintaining them.

Fortunately, Andrew has found us a great new prospect, called the "Heterotic Beast," or HB for short. This is a Ruby-on-Rails package, developed and actively maintained by Jacques Distler, who also is the primary developer for our wiki software Instiki. He is a physicist working in string theory, who John and Andrew have a long relationship with. So we have an ear with him, in case any issues come up, or even feature requests to submit. It's a living project, and that counts for a lot.

Source Text:hoidle

Comments

  • 1.
    edited August 2014

    Since it will be a change, it is important for you to try it out now, before we commit. Please describe any issues that you have with it. I already have a couple, though they are far from being show stoppers. Then we can put all the issues into the balance, and figure out how to proceed.

    I am hoping that we can reach a consensus on this, because I don't know of any other package that meets our needs for markdown and Itex, and that provides such a good support model. But let's see how it goes...

    Source Text:Since it will be a change, it is important for you to try it out now, before we commit. Please describe any issues that you have with it. I already have a couple, though they are far from being show stoppers. Then we can put all the issues into the balance, and figure out how to proceed. I am _hoping_ that we can reach a consensus on this, because I don't know of any other package that meets our needs for markdown and Itex, and that provides such a good support model. But let's see how it goes...
  • 2.

    The first step is for everybody (who wants a say in the matter) to go to the test forum and create an account.

    I posted a few silly threads there.

    Now, go to the link at the top called Signup.

    • Fill in Login name (why not use the same name you have on this forum).

    • Enter a bogus email address, such as zvp@qsd

    • Choose password, and enter it into both fields

    Next, a message will shown to you, saying that you should await the confirmation email. Ignore this. Your account has already been successfully created, with the password you just gave. No email will actually be sent out.

    (Clearly, this is a security issue, so we won't have the real thing running in this mode. But it's good enough for testing purposes.)

    Assuming that we move over to HB, all the messages in the test forum will get wiped clean, and Andrew will copy the data over from the existing forum. So, the conversations there will be transient.

    Source Text:The first step is for everybody (who wants a say in the matter) to go to the [test forum](http://forum.azimuthproject.railsplayground.net/forums/main-forum) and create an account. I posted a few silly threads there. Now, go to the link at the top called Signup. * Fill in Login name (why not use the same name you have on this forum). * Enter a bogus email address, such as zvp@qsd * Choose password, and enter it into both fields Next, a message will shown to you, saying that you should await the confirmation email. _Ignore this_. Your account has already been successfully created, with the password you just gave. No email will actually be sent out. (Clearly, this is a security issue, so we won't have the real thing running in this mode. But it's good enough for testing purposes.) Assuming that we move over to HB, all the messages in the test forum will get wiped clean, and Andrew will copy the data over from the existing forum. So, the conversations there will be transient.
  • 3.

    Next step, everyone please create a thread saying Hello, or something to that effect.

    Make some replies to other people's Hello message. Talk about sports or what have you :)

    We've just got to exercise it!

    Source Text:Next step, everyone please create a thread saying Hello, or something to that effect. Make some replies to other people's Hello message. Talk about sports or what have you :) We've just got to exercise it!
  • 4.

    Glitch 1: The account names all get lowercased.

    As the accounts get created, I will fix this up, by updating the login name in the user table to use the correct case.

    Source Text:Glitch 1: The account names all get lowercased. As the accounts get created, I will fix this up, by updating the login name in the user table to use the correct case.
  • 5.

    Wow, I just noticed something of concern. There's no apparent category system for the posts.

    I will post to the HB forum to ask about this.

    Given the fact that the server is no longer guaranteed, we may be faced with some harder choices than I had anticipated.

    Please go through with the step of creating the account, and making some conversation there. We should gather all the feedback that we can on the package, and send it to Jacques. That will be a way of giving back to him for what he has already given us, and may lead to further improvements in the package that will suit our needs. If I get really inspired, I may contribute to his development efforts.

    One way to get the thing exercised would be to double post all of our discussions about HB to the test forum itself (hint, hint). We should definitely have it here, since this is our place of record, but why not go the extra yard and post these comments to the test forum itself.

    Thanks!

    Source Text:Wow, I just noticed something of concern. There's no apparent category system for the posts. I will post to the HB forum to ask about this. Given the fact that the server is no longer guaranteed, we may be faced with some harder choices than I had anticipated. Please go through with the step of creating the account, and making some conversation there. We should gather all the feedback that we can on the package, and send it to Jacques. That will be a way of giving back to him for what he has already given us, and may lead to further improvements in the package that will suit our needs. If I get really inspired, I may contribute to his development efforts. One way to get the thing exercised would be to double post all of our discussions about HB to the test forum itself (hint, hint). We should definitely have it here, since this is our place of record, but why not go the extra yard and post these comments to the test forum itself. Thanks!
  • 6.
    edited August 2014

    Hi Dave, Great stuff as always. I've posted a new discussion on Urban climate policy which should be a wiki page. One problem, I can't see a button allowing me to edit my post. Perhaps somebody can try replying to it? Cheers

    Source Text:Hi Dave, Great stuff as always. I've posted a new discussion on Urban climate policy which should be a wiki page. One problem, I can't see a button allowing me to edit my post. Perhaps somebody can try replying to it? Cheers
  • 7.
    edited August 2014

    In the new topic page, it has a box for 'name', which would be clearer as 'topic' or 'name of topic'.

    I couldn't preview my post, but (unlike Jim) I did find an edit button. I'd prefer preview, but can live with just edits.

    I do have a "meta-concern" about the HB software, which is that it seems very reliant on one individual, Jacques Distler.

    I made a SVG image there which is cool, and cannot be done here.

    Source Text:In the new topic page, it has a box for 'name', which would be clearer as 'topic' or 'name of topic'. I couldn't preview my post, but (unlike Jim) I did find an edit button. I'd prefer preview, but can live with just edits. I do have a "meta-concern" about the HB software, which is that it seems very reliant on one individual, Jacques Distler. I made a SVG image there which is cool, and cannot be done here.
  • 8.

    Wow, I just noticed something of concern. There’s no apparent category system for the posts.

    I guess you can create multiple sub-forums.

    Source Text:> Wow, I just noticed something of concern. There’s no apparent category system for the posts. I guess you can create multiple sub-forums.
  • 9.

    Tnx for the heads up Graham. I found the edit button.

    Source Text:Tnx for the heads up Graham. I found the edit button.
  • 10.

    Graham wrote:

    I do have a “meta-concern” about the HB software, which is that it seems very reliant on one individual, Jacques Distler.

    I agree that this should be added to the list of Cons for HB.

    I'm predicting that all of our major options will have a significant Con list.

    Let's proceed with the evaluation of HB, noting everything as we go. I started a thread on their forum with our questions, feedback, and "feature requests." (with the caveat that they should take our feature requests with a grain of salt, and not do any work on our behalf alone, since we might not end up adopting it).

    Source Text:Graham wrote: > I do have a “meta-concern” about the HB software, which is that it seems very reliant on one individual, Jacques Distler. I agree that this should be added to the list of Cons for HB. I'm predicting that all of our major options will have a significant Con list. Let's proceed with the evaluation of HB, noting everything as we go. I started a thread on their forum with our questions, feedback, and "feature requests." (with the caveat that they should take our feature requests with a grain of salt, and not do any work on our behalf alone, since we might not end up adopting it).
  • 11.

    On the other hand, he is the chief developer, is responsive, and wants to help us out.

    For instance, when I pointed out the problem with the login names being lowercased, he fixed it on the spot (in the trunk). In this sense, there is a lower barrier to entry for support changes than with a larger project, where features have to be approved by committee or by consensus.

    Source Text:On the other hand, he is the chief developer, is responsive, and wants to help us out. For instance, when I pointed out the problem with the login names being lowercased, he fixed it on the spot (in the trunk). In this sense, there is a lower barrier to entry for support changes than with a larger project, where features have to be approved by committee or by consensus.
  • 12.
    edited August 2014

    Thanks to the people who have tried out the forum so far, which is Andrew, Jim, John, Graham and Paul and myself.

    Since the forum is now running without guaranteed server support, it behooves us to proactive about choosing new forum software and moving over to it. I know we're all busy with other things (myself included), but this could be a case where "the squeaky wheel gets the oil" gets us into trouble. Since the current forum is working really well now, in all likelihood it will shutdown before it squeaks.

    So to anyone who uses the forum and cares about it, at your convenience, please create a login on the new test forum, give it a try, and share your observations.

    Source Text:Thanks to the people who have tried out the forum so far, which is Andrew, Jim, John, Graham and Paul and myself. Since the forum is now running without guaranteed server support, it behooves us to proactive about choosing new forum software and moving over to it. I know we're all busy with other things (myself included), but this could be a case where "the squeaky wheel gets the oil" gets us into trouble. Since the current forum is working really well now, in all likelihood it will shutdown before it squeaks. So to anyone who uses the forum and cares about it, at your convenience, please create a login on the new test forum, give it a try, and share your observations.
  • 13.

    On the test forum, John wrote:

    Sorry to be so forgetful, but will switching to this new forum mean all the old posts get 1) lost, 2) frozen in amber, or 3) ported over here?

    Some people have (foolishly) put a lot of work into putting nice graphics and other material on the old forum, which they should have put onto the wiki.

    The ability to have different categories of posts seems quite important; losing it would be quite a setback. Combined with Graham’s “meta-concern”, it makes me wonder if we can find better software for the forum.

    There are a number of points and questions here, which I will address separately.

    Source Text:On the test forum, John wrote: > Sorry to be so forgetful, but will switching to this new forum mean all the old posts get 1) lost, 2) frozen in amber, or 3) ported over here? > > Some people have (foolishly) put a lot of work into putting nice graphics and other material on the old forum, which they should have put onto the wiki. > > The ability to have different categories of posts seems quite important; losing it would be quite a setback. Combined with Graham’s “meta-concern”, it makes me wonder if we can find better software for the forum. There are a number of points and questions here, which I will address separately.
  • 14.

    John wrote:

    Sorry to be so forgetful, but will switching to this new forum mean all the old posts get 1) lost, 2) frozen in amber, or 3) ported over here?

    Options (1) or (2) would represent a big loss. All along our plan has been to convert the data from the old forum to the new one. Andrew has already made progress in figuring out how to convert the data.

    So let's cross this concern off the map.

    Source Text:John wrote: > Sorry to be so forgetful, but will switching to this new forum mean all the old posts get 1) lost, 2) frozen in amber, or 3) ported over here? Options (1) or (2) would represent a big loss. All along our plan has been to convert the data from the old forum to the new one. Andrew has already made progress in figuring out how to convert the data. So let's cross this concern off the map.
  • 15.
    edited August 2014

    John wrote:

    The ability to have different categories of posts seems quite important; losing it would be quite a setback. Combined with Graham’s “meta-concern”, it makes me wonder if we can find better software for the forum.

    Shortly I will post Jacques response to the question of categories, which gives a hint about the kind of support that he can offer us.

    The support model is a definite factor that we need to weigh carefully. I'll say right now that I have yet to see any option, for a forum that supports itex + markup, that offers a support model that I am truly comfortable with.

    Let's keep gathering information, and compiling our lists of pros and cos. (I already have some more questions about the functioning of the new software, which I'll add to this thread.) Today, the most useful thing would be for everyone with a stake in the forum to exercise it a bit and share their feedback on how it works. After that, we can move on to a deliberation phase.

    Source Text:John wrote: > The ability to have different categories of posts seems quite important; losing it would be quite a setback. Combined with Graham’s “meta-concern”, it makes me wonder if we can find better software for the forum. Shortly I will post Jacques response to the question of categories, which gives a hint about the kind of support that he can offer us. The support model is a definite factor that we need to weigh carefully. I'll say right now that I have yet to see any option, for a forum that supports itex + markup, that offers a support model that I am truly comfortable with. Let's keep gathering information, and compiling our lists of pros and cos. (I already have some more questions about the functioning of the new software, which I'll add to this thread.) Today, the most useful thing would be for everyone with a stake in the forum to exercise it a bit and share their feedback on how it works. After that, we can move on to a deliberation phase.
  • 16.

    Regarding category suport, Jacques Distler wrote:

    Heterotic Beast supports the following hierarchy

    Sites ⊃ Forums ⊃ Topics ⊃ Posts

    Different Sites have different URLs (virtual hosts, in Apache parlance), and different user-lists. For instance, the LHC Forum is another Site, running under the same Heterotic Beast instance as this one.

    The thing closest to Vanilla’s “categories” is Heterotic Beast’s “Forums”. I’m not sure we need both. But we can discuss it.

    So we could, for example have one sub-forum for Technical posts.

    Certainly there is a loss of granularity here. That counts as a negative, but whether or not it is a show-stopper can only be assessed, later, in the context of the assessment of our other options.

    Also it is mitigated by the fact that the lead developer is willing to consider making feature enhancements if we can convince him that they are necessary.

    So, regarding the support issue, we do have the negative that it's just one person. We also have the positive that he is a fantastic developer who we have a direct connection to. Think of how hard it would be to get a feature request into a software project that involves many lead and contributing developers.

    Difficult decisions lie ahead. Today, let's gather more information.

    Source Text:Regarding category suport, Jacques Distler wrote: > Heterotic Beast supports the following hierarchy > > Sites ⊃ Forums ⊃ Topics ⊃ Posts > > Different Sites have different URLs (virtual hosts, in Apache parlance), and different user-lists. For instance, the LHC Forum is another Site, running under the same Heterotic Beast instance as this one. > > The thing closest to Vanilla’s “categories” is Heterotic Beast’s “Forums”. I’m not sure we need both. But we can discuss it. So we could, for example have one sub-forum for Technical posts. Certainly there is a loss of granularity here. That counts as a negative, but whether or not it is a show-stopper can only be assessed, later, in the context of the assessment of our other options. Also it is mitigated by the fact that the lead developer is willing to consider making feature enhancements if we can convince him that they are necessary. So, regarding the support issue, we do have the negative that it's just one person. We also have the positive that he is a fantastic developer who we have a direct connection to. Think of how hard it would be to get a feature request into a software project that involves many lead and contributing developers. Difficult decisions lie ahead. Today, let's gather more information.
  • 17.

    Stepping back, let's consider what is the list of our options. Given that we want the forum to support Itex + Markup, I have seen two options:

    • The current forum, based on a customized version of Vanilla 1.1.10

    • Heterotic Beast, developed and maintained by Jacques Distler (who also is the developer of Instiki)

    Does anybody know of others?

    Source Text:Stepping back, let's consider what is the list of our options. Given that we want the forum to support Itex + Markup, I have seen two options: * The current forum, based on a customized version of Vanilla 1.1.10 * Heterotic Beast, developed and maintained by Jacques Distler (who also is the developer of Instiki) Does anybody know of others?
  • 18.
    edited August 2014

    Regarding the current software, let me summarize the pros and cons, as I see them.

    Pros:

    • It's really great software, and I have no real complaints about it. Functionality is very cool, as is the look and feel.

    • Vanilla is a widely used open-source forum package

    As a user, I wish would could continue on this platform.

    But from an administrative perspective, consider the following drawbacks, which could hit us now, or down the line.

    Cons:

    • Customizations are extensive. Here is a summary of the customizations. See the section called "Converting the nForum to MathML."

    • As of now, the customizations are locking us into a rather old version of Vanilla, along with rather old versions of all of the contributing modules that were modified.

    • There is the question as to whether we could get all those old versions to run in a modern library environment, or, if could be done, how difficult it would be.

    • Even if we could get the old software working on the new installation, it may work today, but fail down the line, when the system administrators update the system libraries.

    • The role is being transition from Andrew to me, but I don't have any knowledge about the internals of this software or MathML.

    • Therefore, without Andrew, we would be effectively supportless with the current software. I'm not saying I wouldn't try to learn about it, but having one amateur in charge of the project, using obsolete versions of the forum code and its contributing modules, is about the worst support model that I can think of.

    Source Text:Regarding the current software, let me summarize the pros and cons, as I see them. Pros: * It's really great software, and I have no real complaints about it. Functionality is very cool, as is the look and feel. * Vanilla is a widely used open-source forum package As a _user_, I wish would could continue on this platform. But from an administrative perspective, consider the following drawbacks, which could hit us now, or down the line. Cons: * Customizations are extensive. Here is a [summary of the customizations](http://www.math.ntnu.no/~stacey/HowDidIDoThat/nLab/). See the section called "Converting the nForum to MathML." * As of now, the customizations are locking us into a rather old version of Vanilla, along with rather old versions of all of the contributing modules that were modified. * There is the question as to whether we could get all those old versions to run in a modern library environment, or, if could be done, how difficult it would be. * Even if we could get the old software working on the new installation, it may work today, but fail down the line, when the system administrators update the system libraries. * The role is being transition from Andrew to me, but I don't have any knowledge about the internals of this software or MathML. * Therefore, without Andrew, we would be effectively supportless with the current software. I'm not saying I wouldn't try to learn about it, but having one amateur in charge of the project, using obsolete versions of the forum code and its contributing modules, is about the worst support model that I can think of.
  • 19.

    Sidenote: In an ideal world, Andrew would have lobbied to get his changes checked into the trunk of Vanilla, and its supporting modules. It deserves to be there.

    Now of course this is easier said than done, given the fact that these projects involve many developers and perspectives, and may or not be well organized. So it could be a full-time job to convert these customizations into a lasting contribution to open source...and I understand that Andrew already has a job :)

    Source Text:Sidenote: In an ideal world, Andrew would have lobbied to get his changes checked into the trunk of Vanilla, and its supporting modules. It deserves to be there. Now of course this is easier said than done, given the fact that these projects involve many developers and perspectives, and may or not be well organized. So it could be a full-time job to convert these customizations into a lasting contribution to open source...and I understand that Andrew already has a job :)
  • 20.

    That's not completely infeasible. The main issue is porting the itex part and making sure that everything is valid XHTML. Most of the plugins that I added (or created or modified) aren't strictly necessary, and those that are probably have equivalents in the current Vanilla set up. Looking down that list, the vast majority were fixes for things that Vanilla ought to have been providing and so are quite likely to be integrated into the new software.

    The main headache of shifting to any new forum software is going to be converting the database. As with HB, the question will be as to how well the current database tables match up with the new ones.

    Source Text:That's not completely infeasible. The main issue is porting the itex part and making sure that everything is valid XHTML. Most of the plugins that I added (or created or modified) aren't strictly necessary, and those that are probably have equivalents in the current Vanilla set up. Looking down that list, the vast majority were fixes for things that Vanilla ought to have been providing and so are quite likely to be integrated into the new software. The main headache of shifting to any new forum software is going to be converting the database. As with HB, the question will be as to how well the current database tables match up with the new ones.
  • 21.
    edited August 2014

    Intriguing possibility. Let's think this through, to see if it can actually be feasible. There are two main aspects to this:

    • Getting it working with the current version of Vanilla. This is a technical matter, and your comments above are encouraging.

    • Getting it working as a sustainable, ongoing solution for us, in the face of the inevitable changes in the Vanilla code, and the iTex and XHTML specifications. This would mean having your work adopted by the Vanilla team, which is a social consideration that of course can't be addressed without communicating with them.

    Here is a link to the Vanilla developers forum. One idea would be for you to float your idea by them, and point them to nice examples on the nForum or the Azimuth Forum.

    I see that the development is sponsored by a company, who hires developers. (Not sure what your plans in life are, but with these credentials you may be able to land a good job with them. You could position yourself as their math formatting specialist.) If they control the trunk, then it would make sense to communicate with them. Research needed...

    In the absence of it being truly adopted by them, I'd be concerned about the ongoing maintenance. For instance, suppose you wrote a user-contributed module that works nicely. Then what happens (1) when the module interface evolves over time, and we want to upgrade to the latest version of Vanilla, or (2) the XHTML or iTex standards evolve, and your module is no longer compliant, doesn't make use of important new features in the standard.

    Source Text:Intriguing possibility. Let's think this through, to see if it can actually be feasible. There are two main aspects to this: * Getting it working with the current version of Vanilla. This is a technical matter, and your comments above are encouraging. * Getting it working as a sustainable, ongoing solution for us, in the face of the inevitable changes in the Vanilla code, and the iTex and XHTML specifications. This would mean having your work _adopted_ by the Vanilla team, which is a social consideration that of course can't be addressed without communicating with them. Here is a link to the [Vanilla developers forum](http://vanillaforums.org/categories/developers). One idea would be for you to float your idea by them, and point them to nice examples on the nForum or the Azimuth Forum. I see that the development is sponsored by a company, who hires developers. (Not sure what your plans in life are, but with these credentials you may be able to land a good job with them. You could position yourself as their math formatting specialist.) If they control the trunk, then it would make sense to communicate with them. Research needed... In the absence of it being truly adopted by them, I'd be concerned about the ongoing maintenance. For instance, suppose you wrote a user-contributed module that works nicely. Then what happens (1) when the module interface evolves over time, and we want to upgrade to the latest version of Vanilla, or (2) the XHTML or iTex standards evolve, and your module is no longer compliant, doesn't make use of important new features in the standard.
  • 22.
    edited August 2014

    Overall goal would be to drum up some interest in their community for your math support -- so that you and we would not be the only ones with a stake in the matter.

    Just as a point of reference, here is a commercial package that provides TeX support for Vanilla. They don't have a Linux version.

    Source Text:Overall goal would be to drum up some interest in their community for your math support -- so that you and we would not be the only ones with a stake in the matter. Just as a point of reference, here is a [commercial package](http://www.dessci.com/en/products/mathtype/works_with.asp#!target=vanilla_forums) that provides TeX support for Vanilla. They don't have a Linux version.
  • 23.

    Stepping back, let’s consider what is the list of our options. Given that we want the forum to support Itex + Markup, I have seen two options:

    Is itex a hard requirement given the existence of MathJax and asciimath?

    Source Text:> Stepping back, let’s consider what is the list of our options. Given that we want the forum to support Itex + Markup, I have seen two options: Is itex a hard requirement given the existence of MathJax and asciimath?
  • 24.

    One simple reason would be backwards compatibility.

    Another is that even with MathJaX available, it works best (which partly means fastest) if it is given MathML to work with, not it's LaTeX-based input. So there should be some sort of converter to MathML and itex is one of the best, if not the best.

    Source Text:One simple reason would be backwards compatibility. Another is that even with MathJaX available, it works best (which partly means fastest) if it is given MathML to work with, not it's LaTeX-based input. So there should be some sort of converter to MathML and itex is one of the best, if not the best.
  • 25.

    David wrote:

    So we could, for example have one sub-forum for Technical posts.

    Do you mean one so-called "Forum", according to Jacques' scheme here?

    Sites ⊃ Forums ⊃ Topics ⊃ Posts

    Can we create so-called "Forums" of this sort at will?

    Source Text:David wrote: > So we could, for example have one sub-forum for Technical posts. Do you mean one so-called "Forum", according to Jacques' scheme here? > Sites ⊃ Forums ⊃ Topics ⊃ Posts Can we create so-called "Forums" of this sort at will?
  • 26.

    Yes, that's what I meant to say.

    Source Text:Yes, that's what I meant to say.
  • 27.
    edited September 2014

    John, given that, how much does that change your assessment of the need for Vanilla style categories?

    When I think about it, though it's nice to know that I could do so, I never actually filter the posts by category. The main value that I get out of categories is the act of thinking about what category to put a post in, and seeing the category in the header of the post is informative.

    As for the visibility control, well we're leaning towards making everything visible anyway. If we still want Strategy to be not visible, we could make a separate Forum for it.

    Trouble with having lots of Forums -- unless there is another mechanism I don't know about -- is that one would have to visit them all.

    Source Text:John, given that, how much does that change your assessment of the need for Vanilla style categories? When I think about it, though it's nice to know that I could do so, I never _actually_ filter the posts by category. The main value that I get out of categories is the act of thinking about what category to put a post in, and seeing the category in the header of the post is informative. As for the visibility control, well we're leaning towards making everything visible anyway. If we still want Strategy to be not visible, we could make a separate Forum for it. Trouble with having lots of Forums -- unless there is another mechanism I don't know about -- is that one would have to visit them all.
  • 28.

    Clarification, the administrator can create forums at will, not the general users

    Source Text:Clarification, the administrator can create forums at will, not the general users
  • 29.

    It would be good to clarify where things stand on the issue of the forum conversion.

    So far I see the following options:

    1. Convert to Jacques' forum software, which is what this thread is about

    2. Try to get Andrew's changes checked into the Vanilla trunk, and adopted by the Vanilla developer community

    3. Keep going on the current server that the forum is running under now

    Option (2) would be great if we could pull it off, but it means more development work, and it does not sound either easy or guaranteed that we will succeed in convincing some other developers, who we don't even know, to accept the updated version of Andrew's work. So there's quite a bit of risk here, in terms of loss of time, and the possibility of not succeeding.

    Jacques' software looks to be a plausible option. As pointed out, there is an issue here with him being a one-man developer team. This is partially mitigated -- but only partially! -- by the following. He is quite helpful, and open to adding new features, if we can justify their need. And he is a great one-man team: in fact, we've been relying upon him for support of Instiki, and I have not been hearing any complaints about that software.

    But, all that notwithstanding, Jacques forum software can't really compete in terms of its overall usability, feature set, look-and-feel, and support level, with a forum software that has a large developer base and user community, like Vanilla.

    I myself like Vanilla better (which we are using right now). By the way, Vanilla now comes with Markdown built into it.

    What has happened is that we've come to take for granted the great combination of the general Vanilla package the ITex capability that Andrew has added to it.

    But if the current server setup is not guaranteed to remain in operation, and we don't come up with some other alternatives, we may unfortunately end up having to choose between general quality and the feature of latex support.

    Andrew, what are the risks if we just punt for now, and keep things as they are. What relationships do you have with the people who are running the server, and what commitments have they given you in terms of continuation of service, and notice time if it is to become terminated?

    Source Text:It would be good to clarify where things stand on the issue of the forum conversion. So far I see the following options: 1. Convert to Jacques' forum software, which is what this thread is about 2. Try to get Andrew's changes checked into the Vanilla trunk, and adopted by the Vanilla developer community 3. Keep going on the current server that the forum is running under now Option (2) would be great if we could pull it off, but it means more development work, and it does not sound either easy or guaranteed that we will succeed in convincing some other developers, who we don't even know, to accept the updated version of Andrew's work. So there's quite a bit of risk here, in terms of loss of time, and the possibility of not succeeding. Jacques' software looks to be a plausible option. As pointed out, there is an issue here with him being a one-man developer team. This is partially mitigated -- but only partially! -- by the following. He is quite helpful, and open to adding new features, if we can justify their need. And he is a _great_ one-man team: in fact, we've been relying upon him for support of Instiki, and I have not been hearing any complaints about that software. But, all that notwithstanding, Jacques forum software can't really compete in terms of its overall usability, feature set, look-and-feel, and support level, with a forum software that has a large developer base and user community, like Vanilla. I myself _like_ Vanilla better (which we are using right now). By the way, Vanilla now comes with Markdown built into it. What has happened is that we've come to take for granted the great combination of the general Vanilla package the ITex capability that Andrew has added to it. But if the current server setup is not guaranteed to remain in operation, and we don't come up with some other alternatives, we may unfortunately end up having to choose between general quality and the feature of latex support. Andrew, what are the risks if we just punt for now, and keep things as they are. What relationships do you have with the people who are running the server, and what commitments have they given you in terms of continuation of service, and notice time if it is to become terminated?
  • 30.

    I don't think there are major risks in keeping things as they are for now. The impression that I was given was that the server the forum is on would be allowed to keep running until the point where the sysadmins had to step in to do some maintenance, at which point it would be shut down. Given that they haven't had to do anything in the years since it was set up, it should be good for a while yet.

    Source Text:I don't think there are major risks in keeping things as they are for now. The impression that I was given was that the server the forum is on would be allowed to keep running until the point where the sysadmins had to step in to do some maintenance, at which point it would be shut down. Given that they haven't had to do anything in the years since it was set up, it should be good for a while yet.
  • 31.

    Good to know that there's no cutoff date, but unsettling to think a Poisson event (as it were) could knock us out.

    If we simply wait and move on to other things, we face a two-fold risk: (1) loss of data if the sysadmins, who it sounds like you no longer have close ties to -- correct me if I'm wrong -- just scrap the whole thing in a summary fashion, and (2) the loss of continuity and confusion that would result if we then start trying to figure out where and how to port the data. I myself am "paranoid" about data loss, and even if my best friend tells me that he's backing up my data, or that, just in case he has to delete my virtual machine, I will be able to get a copy of the data -- I won't sleep well until it's secured in a way that either have control over the data, or can verify that the backups are actually working.

    Conclusion: we shouldn't become complacent about finding a long term solution, and shouldn't lose this opportunity that we have now to research possibilities, try things out, and possibly make some difficult choices.

    I'm particularly eager to get the forum data under our control, and have it mirrored along with the wiki data, which I am now mirroring.

    To keep things moving along I propose that we investigate the track of applying your changes to a modern version of Vanilla. The first step would be for me to install it, and poke around to see how it works and what the module interface is. Then we can make an assessment of how cleanly we could make a module or patch. If it looks doable in a clean way, then we can do it. Then we'd have something to work with, and something concrete to try to get contributed into the main Vanilla distribution.

    Andrew does this sound reasonable. If I get this started, would you be up for looking into this with me?

    Thanks

    Source Text:Good to know that there's no cutoff date, but unsettling to think a Poisson event (as it were) could knock us out. If we simply wait and move on to other things, we face a two-fold risk: (1) loss of data if the sysadmins, who it sounds like you no longer have close ties to -- correct me if I'm wrong -- just scrap the whole thing in a summary fashion, and (2) the loss of continuity and confusion that would result if we _then_ start trying to figure out where and how to port the data. I myself am "paranoid" about data loss, and even if my best friend tells me that he's backing up my data, or that, just in case he has to delete my virtual machine, I will be able to get a copy of the data -- I won't sleep well until it's secured in a way that either have control over the data, or can verify that the backups are actually working. Conclusion: we shouldn't become complacent about finding a long term solution, and shouldn't lose this opportunity that we have now to research possibilities, try things out, and possibly make some difficult choices. I'm particularly eager to get the forum data under our control, and have it mirrored along with the wiki data, which I am now mirroring. To keep things moving along I propose that we investigate the track of applying your changes to a modern version of Vanilla. The first step would be for me to install it, and poke around to see how it works and what the module interface is. Then we can make an assessment of how cleanly we could make a module or patch. If it looks doable in a clean way, then we can do it. Then we'd have something to work with, and something concrete to try to get contributed into the main Vanilla distribution. Andrew does this sound reasonable. If I get this started, would you be up for looking into this with me? Thanks
  • 32.

    I had a play around with Vanilla 2 the other night and got further than I expected. My current hang-up is the HTML sanitiser (HtmLawed) which doesn't allow MathML tags. There's a rumour on the web of a version that does but I haven't been able to track it down.

    Source Text:I had a play around with Vanilla 2 the other night and got further than I expected. My current hang-up is the HTML sanitiser (HtmLawed) which doesn't allow MathML tags. There's a rumour on the web of a version that does but I haven't been able to track it down.
  • 33.
    edited September 2014

    Well you got further than me. I installed Vanilla 2.1.3 on our server, it can be accessed through vanilla.azimuthproject.railsplayground.net. Not much to see there, since it's barely configured. (For some reason, the installer put it in a vanilla subfolder of the Instiki public directory.)

    Here's a post from 2010 from a guy who was trying to get Latex support to work in Vanilla. He's also complaining about HtmLawed.

    Source Text:Well you got further than me. I installed Vanilla 2.1.3 on our server, it can be accessed through vanilla.azimuthproject.railsplayground.net. Not much to see there, since it's barely configured. (For some reason, the installer put it in a vanilla subfolder of the Instiki public directory.) Here's a [post from 2010](http://vanillaforums.org/discussion/13640/htmlawed-and-other-input-format) from a guy who was trying to get Latex support to work in Vanilla. He's also complaining about HtmLawed.
  • 34.
    edited September 2014

    David - I'm sorry to have been so inattentive to this issue. I expect whatever you do will be best, but I should weigh in nonetheless.

    Like you, I think there are strong advantages to working with free software that's popular enough that a large team of people keeps improving it. Andrew's software works great, and we've come to rely on his help, but now he wants to quit supporting it, and this puts us in a situation where we have to worry and perhaps do a bunch of technical work. Jacques Distler is older than Andrew, born in 1961 like me, so he too will someday quit supporting his software.

    So, I think what you're doing is great: trying to move us over to a more popular setup, namely Vanilla. What are the deficiencies of Vanilla as it exists?

    1) It doesn't do LaTeX in some nice way. Can it do it in some acceptable way, like using MathJax? Andrew and Jacques are both quite vehement in claiming that's not good enough, but I have a different set of values. To me MathJax seems good enough; even crappy math rendering like what we currently have on our Wordpress blog is good enough.

    2) Is it hard or impossible to port the forum posts we have now to Vanilla until we make changes to Vanilla?

    Source Text:David - I'm sorry to have been so inattentive to this issue. I expect whatever you do will be best, but I should weigh in nonetheless. Like you, I think there are strong advantages to working with free software that's popular enough that a large team of people keeps improving it. Andrew's software works great, and we've come to rely on his help, but now he wants to quit supporting it, and this puts us in a situation where we have to worry and perhaps do a bunch of technical work. Jacques Distler is older than Andrew, born in 1961 like me, so he too will someday quit supporting his software. So, I think what you're doing is great: trying to move us over to a more popular setup, namely Vanilla. What are the deficiencies of Vanilla as it exists? 1) It doesn't do LaTeX in some nice way. Can it do it in some acceptable way, like using MathJax? Andrew and Jacques are both quite vehement in claiming that's not good enough, but I have a different set of values. To me MathJax seems good enough; even crappy math rendering like what we currently have on our Wordpress blog is good enough. 2) Is it hard or impossible to port the forum posts we have now to Vanilla until we make changes to Vanilla?
  • 35.

    Thanks for pointing out the MathJax option, it definitely sounds like it merits investigation. Did a web search, and I did find people talking about using MathJax under Vanilla. There is at least one MathJax plugin, user-contributed.

    I agree that it may be worthwhile for us to lower our standards for rendering quality, in order to reach a supportable solution.

    Personal opinion: I certainly care about quality rendering on the wiki and the blog, but to me the forum is somewhat like the back of a napkin, so I don't mind a lower standard. In fact, when I can get away with it for simple expressions, I prefer to dash things off in a textual manner, like Pressure(t) = Energy(t) / Volume(t).

    One thing to consider is whether our existing posts will be parsed by a MathJax plugin, or whether we would have to run a preprocessing script to convert them to the syntax of the plugin. Another question in my mind is what percentage of Azimuth posts actually use Latex. Suppose for the sake of argument, that 5% of our messages used math, and there was a glitch in a MathJax plugin that caused 5% of these not to render properly, in particular some messages from 2008? Then could we live with this loss, in order to obtain the gain of a well-supportable solution?

    You see of course that this is a leading question, intended to whittle down our expectations, before we actually look at the MathJax output :)

    Source Text:Thanks for pointing out the MathJax option, it definitely sounds like it merits investigation. Did a web search, and I did find people talking about using MathJax under Vanilla. There is at least one MathJax plugin, user-contributed. I agree that it may be worthwhile for us to lower our standards for rendering quality, in order to reach a supportable solution. Personal opinion: I certainly care about quality rendering on the wiki and the blog, but to me the forum is somewhat like the back of a napkin, so I don't mind a lower standard. In fact, when I can get away with it for simple expressions, I prefer to dash things off in a textual manner, like Pressure(t) = Energy(t) / Volume(t). One thing to consider is whether our existing posts will be parsed by a MathJax plugin, or whether we would have to run a preprocessing script to convert them to the syntax of the plugin. Another question in my mind is what percentage of Azimuth posts actually use Latex. Suppose for the sake of argument, that 5% of our messages used math, and there was a glitch in a MathJax plugin that caused 5% of these not to render properly, in particular some messages from 2008? Then could we live with this loss, in order to obtain the gain of a well-supportable solution? You see of course that this is a leading question, intended to whittle down our expectations, before we actually look at the MathJax output :)
  • 36.
    edited September 2014

    Re: The conversion of posts from our current forum to the new version of Vanilla.

    The first question doesn't pertain to latex at all. This is the question of whether there has been any substantial change in the structure of the database tables, from the Azimuth verison of Vanilla, to the latest one. If not, we may be able to simply export the data from the current database, and import it into the new Vanilla database. If there have been changes, then we may need to write some scripts to massage the data into the new structure.

    Once the data is loaded, then, we should see the posts. Without a MathJax plugin, they would show up, but the Laxtex wouldn't render, it would show as the source text.

    Source Text:Re: The conversion of posts from our current forum to the new version of Vanilla. The first question doesn't pertain to latex at all. This is the question of whether there has been any substantial change in the structure of the database tables, from the Azimuth verison of Vanilla, to the latest one. If not, we may be able to simply export the data from the current database, and import it into the new Vanilla database. If there have been changes, then we may need to write some scripts to massage the data into the new structure. Once the data is loaded, then, we should see the posts. Without a MathJax plugin, they would show up, but the Laxtex wouldn't render, it would show as the source text.
  • 37.

    David wrote:

    I agree that it may be worthwhile for us to lower our standards for rendering quality, in order to reach a supportable solution.

    Good! Such things as 1) being able to easily manage lots of conversations of different types and 2) link back and forth between the forum and the wiki seem more important than rendering quality to me.

    Suppose for the sake of argument, that 5% of our messages used math, and there was a glitch in a MathJax plugin that caused 5% of these not to render properly, in particular some messages from 2008? Then could we live with this loss, in order to obtain the gain of a well-supportable solution?

    If 1) people can see the source code and guess what's meant, 2) you and I, as administrators, can go back and fix other people's posts when we want, and 3) everyone can fix their own posts if they want, then this seem perfectly acceptable. If just some of 1)-3) hold, well, it's still tolerable.

    I imagine someone occasionally saying "what the heck were you saying back there", and then we fix it up.

    Source Text:David wrote: > I agree that it may be worthwhile for us to lower our standards for rendering quality, in order to reach a supportable solution. Good! Such things as 1) being able to easily manage lots of conversations of different types and 2) link back and forth between the forum and the wiki seem more important than rendering quality to me. > Suppose for the sake of argument, that 5% of our messages used math, and there was a glitch in a MathJax plugin that caused 5% of these not to render properly, in particular some messages from 2008? Then could we live with this loss, in order to obtain the gain of a well-supportable solution? If 1) people can see the source code and guess what's meant, 2) you and I, as administrators, can go back and fix other people's posts when we want, and 3) everyone can fix their own posts if they want, then this seem perfectly acceptable. If just some of 1)-3) hold, well, it's still tolerable. I imagine someone occasionally saying "what the heck were you saying back there", and then we fix it up.
  • 38.

    So we have some leads!

    Here are some steps that we can take:

    1) Configure the Vanilla 2.1.3 that I installed on vanilla.azimuthproject.railsplayground.net.

    2) Give it a nice looking theme, which at least stays close to the spirit of what we have now

    3) Take a backup of the forum data, and copy it over to the new server

    4) Look into loading the data into the new Vanilla database

    5) Explore math rendering options. MathJax, or porting Andrew's changes

    All else being equal, I would prefer the MathJax approach, because it doesn't call for custom support (either from Andrew or from me, who has no expertise with the internals of math formatting software). The question would be, how much of a loss of quality or elegance could we live with. Let's see what we find...

    Source Text:So we have some leads! Here are some steps that we can take: 1) Configure the Vanilla 2.1.3 that I installed on vanilla.azimuthproject.railsplayground.net. 2) Give it a nice looking theme, which at least stays close to the spirit of what we have now 3) Take a backup of the forum data, and copy it over to the new server 4) Look into loading the data into the new Vanilla database 5) Explore math rendering options. MathJax, or porting Andrew's changes All else being equal, I would prefer the MathJax approach, because it doesn't call for custom support (either from Andrew or from me, who has no expertise with the internals of math formatting software). The question would be, how much of a loss of quality or elegance could we live with. Let's see what we find...
  • 39.

    Steps (1) - (4) will be needed for any Vanilla-based solution.

    I think we should do step (3) as soon as possible, given that as of now we don't have any backup copy of the forum data. Andrew I'll send you an email to discuss this. Thanks.

    Source Text:Steps (1) - (4) will be needed for any Vanilla-based solution. I think we should do step (3) as soon as possible, given that as of now we don't have any backup copy of the forum data. Andrew I'll send you an email to discuss this. Thanks.
  • 40.

    Regarding steps 1, 2 and 3, Andrew either you or I, or both of us can work on them. You'll be able to do it many times faster than me, which would be cool, but if you're busy I'm happy to pick up on it. I will send you the details about the installation.

    Source Text:Regarding steps 1, 2 and 3, Andrew either you or I, or both of us can work on them. You'll be able to do it many times faster than me, which would be cool, but if you're busy I'm happy to pick up on it. I will send you the details about the installation.
  • 41.

    MathJaX isn't a silver bullet here. You'd still need to modify the (Markdown) parser so that the maths part didn't get processed, otherwise special characters like * would get interpreted as Markdown markup.

    I've managed to get Vanilla 2 working with iTeX. I needed to modify the HtmLawed plugin to allow the maths tags through and I haven't added all of them yet. I copied the Markdown+iTeX plugin straight from this forum, which will mean that the links to the wiki will work as well (though I've ditched the configuration options since that's all changed).

    Source Text:MathJaX isn't a silver bullet here. You'd still need to modify the (Markdown) parser so that the maths part didn't get processed, otherwise special characters like `*` would get interpreted as Markdown markup. I've managed to get Vanilla 2 working with iTeX. I needed to modify the HtmLawed plugin to allow the maths tags through and I haven't added all of them yet. I copied the Markdown+iTeX plugin straight from this forum, which will mean that the links to the wiki will work as well (though I've ditched the configuration options since that's all changed).
  • 42.
    edited September 2014

    Cool, I'm looking forward to seeing how this will look with the new Vanilla.

    We can consider MathJax as a backup plan, in case we get stuck.

    Thanks Andrew.

    Source Text:Cool, I'm looking forward to seeing how this will look with the new Vanilla. We can consider MathJax as a backup plan, in case we get stuck. Thanks Andrew.
  • 43.

    I think we should do step (3) as soon as possible, given that as of now we don’t have any backup copy of the forum data.

    What's with the old backup copies?

    Source Text:>I think we should do step (3) as soon as possible, given that as of now we don’t have any backup copy of the forum data. What's with the old backup copies?
Sign In or Register to comment.