As long as the new site still has a medal system I’m all for it.
The $5 Linode happily runs puppygames.net, and for some reason, is way faster than the $150 Rackspace VPS.
We want to go to another host because of the host being unreliable?
Or are the problems with reliability related to the SMF software?
The last few days, JGO has been stable, yes?
But I still think the forum categories need a redo.
If @princec sets up an Apache server and MYSQL as offered, will the current SMF run on it?
Maybe the thing to do is install the new forum as a sub-domain, at least for now, while we set it up with the proper categories and the like, and see if people like it? If/when it gets a thumbs up, we can give it the java-gaming.org url and make the SMF a read-only sub-domain?
Or just set up the sub-domain and new forum for Java9+ related Java-gaming?
@SHC has written code to import much of JGO into Flarum, but has some issues with it that I am having trouble grasping.
But I’m not clear if we really need to follow through on that, if we can manage to have a “legacy” and a current version coexisting. (I’m not wanting to install apache, PHP, composer, flarum, and learn how to code PHP and run the various components, that is a lot of work to take on for a task that may be moot.)
Meanwhile, Flarum just released version 0.1.0-beta.9 a couple days ago, for what that is worth.
What is the a reason some are preferring Flarum over SMF2 as new hosting software?
Do either/both offer the ability to customize things like we’ve done (e.g., profile pictures, a form of “appreciations”)?
Totally happy if we just archive this entire forum under /archive, then start completely from scratch with brand new forums and get everyone interested to register again. There are only a handful of active posters and a slightly larger number of regular lurkers for whom this will be the most minutest issue to do, and then we can go from there and see about asking Riven to transfer the domains.
First things first though, choose forums software. I can set up a $5 Linode with SSH access for someone, and that’ll be my contribution. Whoever has the SSH access will have root, and will be able to do what they want regarding setting up Apache, MySQL, forums, etc.
Side question: are there any forum software out there that doesn’t use PHP and MySQL? Because frankly both suck in various facepalming ways… if I had my way it’d be running as JSPs on Jetty with a Postgres back-end.
Makes sense to me to use Java-based programs for forum and servers. “Eat your own dog food.”
If we went to JForum2, would we use Tomcat? I have actually used Tomcat before, though only for dev, locally.
IDK about switching from MySQL to PostgreSQL. It adds another opportunity for surprises/unknowns to enter the migration process. @princec: what would be the benefits that would make adding a bit more risk/effort worth it? MySQL has a rep of being faster and easier to use and code (justified?).
SHC made a good argument in an email to me for migrating the entire forum vs. leaving two running. Changing it would require some coding (e.g., to make it read-only, or add links to the new db). And it is reportedly already pretty iffy in terms of being able to make ANY changes whatsoever.
Postgres is better in most ways than MySQL… I’ve used both pretty extensively. But I couldn’t care less tbh Jforum2 seems to work with whatever database engine you point it at.
As for migration… what is the actual point if we want the old site archived? Just create a brand new site, start from scratch. Simply disable all the logins in the old forum and delete the user signup page.
wrt Tomcat - I’ve used it, it’s a PITA, but it works. Jetty will also work, possibly less hassle. Either of them can be used exclusively - no need for Apache, which might be another blessing. I’ve just spent a few days configuring a brand new Apache installation and it’s quite a crapfest.
@philfrei: Nope that is effectively the opposite to what i suggested.
My suggestion was that if the forum was somehow renewed, the old forum should not be moved to a new subdirectory like “/legacy”;
doing so would mean that you could no longer reach any thread or post from the current forum under its old exact URL, meaning that almost every link pointing to one of those (be it cross-references between old forum posts or even links from other sites pointing to this site) would no longer reach its destination, thus: link rot.
My suggestion was to take any action only while keeping that in mind and trying to work around it.
Specifically, leaving the old site where it is, putting a new site in its own location and only forwarding from the old to the new one.
And then again I would like to note the absence of any contribution to this from Riven who, AFAIK is the only person ultimately having a say in this.
@Drenius Thanks for setting me straight on that. That makes another good argument for porting over all the existing posts to a new forum, assuming we can maintain urls while doing so.
I’m trying to create some sort of map with pending decisions and actions, based on what I’ve learned from this discussion so far. Given that we are doing this by something even less organized than a committee, it is going to take some time to hash things out. But I think it would be good to push forward, even with the site staying accessible this last couple weeks, rather than waiting for it to crash again.
The numbering is mostly to make it easier to refer to specific steps. My understanding is full of holes, and the outline is VERY rough. I am hoping for corrections and fill-ins. Also, ideas for where to put this worksheet and make it more readable. Maybe give it a topic on Discussions/Gaming Wiki, so anyone can edit? Or let me know things and I will edit here on this page.
- princec has offered to pay for and “administer” a hosting site (recommends Linode, under $10/mo)
Correct me if I am wrong: administration refers to making sure the host gets paid on time? Might also include setting up the server (if Apache or Jetty–princec’s preference over Tomcat).
I am unclear on the administration/cost required for retaining the url (“java-gaming.org”). DNS registration costs something like $20/year?
We can get a “dump” of the current forum when the time comes.
I’m not clear how a new snapshot might be created. Perhaps a request is made to @Riven? Perhaps the code to do this is already written and just needs to be executed.
We can create a MySQL database from the forum dump. It is a single command (LOAD SQL) executed at a command shell level. I’ve run it on the snapshot of the forum taken a year ago and it works fine.
Decision 1: forum software.
Two main candidates:
Some pros and cons:
a1) Flarum can use the “same” back end as currently exists, which is PHP based and uses Apache.
a2) Code for importing has been given a good start by @SHC on github. Due to limitations of SQL, importing is very slow (one record at a time) and the code does not complete even after 6 hours of running. But there are two strategies that can used to significantly improve performance. https://dev.mysql.com/doc/refman/5.6/en/insert-optimization.html One is to take advantage of a MySQL feature: the ability to INSERT multiple value lists in a single statement. An example. I think someone who already knows PHP and has a dev environment for it would be able make this code change pretty easily from the existing github project. (I do not know PHP, do not have dev environment for it.)
The second strategy is to import from a delimited text file (using LOAD DATA). The text file could either be created from the MySQL database created from the forum dump (using SELECT…INTO), or perhaps the code which does the forum dump could be tweaked to create a delimited text file that could be used with “LOAD DATA” rather than “LOAD SQL”.
a3) Flarum is not stable yet. A new beta (version 9) was created a couple weeks ago. They are not yet to 1.0. The code on github imports to beta version 7. Unclear what modifications would have to be made or if the makers of Flarum have provided reliable upgrade path. I think we did spot material for upgrading from one beta to the next. One just doesn’t know how well it will work.
b1) JForum2 is Java-based. A benefit is that we would have more candidates to help out, do customizations, than with a PHP-based back end.
b2) I don’t think anyone has scoped out the initialization tasks involved with JForum2 yet, or the coding we would need to do the importing. It is possible that there exists some SMF to JForum2 code already written–this would have to be researched.
b3) JForum2 can run on a Tomcat or Jetty back end.
A feature comparison of the two Forums would possibly be helpful.
IDK. My weak attempt here at project management, is it better than nothing or is it inhibiting people with more initiative and ability?
Maybe I’ll just download Jetty [DONE] and JForum2 to my laptop and see if I can set them up [Jetty’s “Demo-base” working–that was too easy] and then set up a replica of the current forum sections. I’d rather do that than learn PHP. I hope dealing with Jetty isn’t as much of a learning curve as it was to get Tomcat running (something I did for myself several years ago). So put me down for doing some investigation on 4b2.
I agree about trying to keep the old forum at the same URL, it’ll make finding legacy posts that much more likely to work.
The new forum could be at www.java-gaming.org/forums for example, and the index page could redirect there.
As far as administration is concerned, I’m proposing to pay for the Linode, set up SSH, and then let someone else loose on it to do the dirty work. The domains is owned by Riven and you’ll have to ask him nicely to point it at the new server when everyone’s good and ready.
Failing all this… javagaming.org (the original url!) redirects to java-gaming.org, and it wouldn’t be entirely beyond the realms of good sense to take that domain instead, leaving java-gaming.org exactly as it is, and have www.javagaming.org just be the new forum. Which is arguably much simpler than all of the other solutions as it doesn’t involve any migration of the old stuff, just leave it be.
Sorry if I didnt follow the last developments, but has Discourse been ruled out?
I don’t think it has, go for it.
I didn’t rule it out. I have mostly been just responding to what has been posted on this thread, and this is the first time it was mentioned here. This is an attribute of action by forum-committee where participation is intermittent.
Here is a proposal. Set up some new forums and see what they look like before making a choice.
I’ve started working on JForum2.
@elect, any interest in taking the lead in setting up a Discourse example?
@SHC, do you have time to continue on the Flarum version? (I’ve been meaning to check in with you if my research on MySQL “batching” was helpful or not.)
We have last year’s snapshot to use as a basis.
I’m not sure how we set up three test sites. Maybe it can be done at a subdomain of jgo? If not, maybe a subdomain at my url (just for test comparisons). I’m using Startlogic. I’m not clear about running server software there. I’ll see what I can find out. Maybe I should try switching over to a linode account. Startlogic has boosted the per month rate a fair bit since I started with them.
YES – the notion of javagaming.org or java-gaming.org/forums with redirect sounds good. Main issue not covered by this is concern about existing forum software “breaking” again?
I could see having a pinned link to the “archive” version with each topic.
Some years ago I setup a Discourse forum on DigitalOcean trying to push the jogamp community into something modern.
I recall I was relieved how much easy it was, I have nothing to believe something has changed in the meanwhile.
I think Discourse is one of the best (if not the) software forum out there, here a list of features.
Among those I love the clean interface, dynamic notifications, born mobile, spam protection, two-factor authentication and 100% open source.
I also like some other small features, which give an idea about the level of attention the authors (which are the same of StackOverflow) put into, like:
- automatic collapsing for huge conversations
- automatic sketchs saving (as StackOverflow)
- automatic detection if you repost a link in the same discussion
- dynamic suggestion of already open threads as long as you type (as StackOverflow)
To sum up, I’d love to give my contribution, how can we go on?
Ok, I’m going to set up a minimum-size Linode this afternoon then.
New server created - PM’ed you on JGO with details.
I’m having troubles getting JForum2 to run on Jetty. Am noticing that only Tomcat is listed in system requirements. There was a document from several years ago on how to run on Jetty but it has “disappeared”. I thought I saw something somewhere that said it would work but am now not seeing it.
I’m waiting for the JForum help to indicate whether it will run on Jetty or not. It it requires Tomcat, so be it.
Discourse does look nice. Was just checking out their github site.
I found this ancient list of open source Java-built forum software, for what it is worth. I’d kind of prefer to have something that is built with Java. But if it can’t compete, I’m not going to object.
I think it would be better to take some working software instead of preffering Java-based forum software.
This forum is about java-gaming and not about java-forums
I’d prefer to have staable and efficent forum against java-based forum.
Well, both JForum2 and Discourse work. So: let’s try both and see what we like before progressing. Over to you Phil.
- given lwjgl and jgo are both relatively low traffic, it’d make sense to merge them together in order to concentrate everything in one place and avoid dispersion. Also this mean, theoretical, lower maintenance… two birds one stone. Apostolos and Spasi seem to agree and share the same point of view
I’ll mirror the lwjgl forum into a jgo subforum.
- @philfrei, why don’t you join us on slack? we could coordinate us faster and easier