Please find on github a quick write-up:
1.0 final (https://github.com/Elkarte/ElkArte/issues/2)
These (roughly written) targets are subject to clarify, which is probably why I am opening this discussion here as well.
As a side note... It may be preferable however if we keep things more on github than a test forum? At least, please, do make use of our github home, a couple of us are using it all the time and information flow is easier there. (from code reviews and PRs to the start-up wiki and issues trackers.)
/me also knows that Emanuele likes deleting topics... :-X
Please, add your pet hates/loves, ask your questions, lets clarify this a little better. We haven't hurried with roadmap write-up, only quick exchanges have sufficed (and probably may still), however perhaps it's better to make sure we know the big picture and are in line here?
Oh, and... I should add. My personal aim is to have the intentions finished enough for a good beta by March... The March in like 3 months. I is not joking... (that I know of, yet) and that's because I was thinking to keep the indicative goal for stable for October, and considering that I don't know what will happen elsewhere.
Fixed link to the "new" repo. :P
BTW, so, so, so, in practical terms that even someone like me can understand, what is missing now? O:-)
From the issue, I see parse_bbc rewriting, but I wonder if it make sense to rewrite it now (it has been like that for years one more doesn't change much :P).
Refactoring? What's the current status? What is missing for 1.0 "beta"?
Theme? After the latest commits I have no idea of the current state (not that I had any before lol)
Is it reasonable to see a beta in September?
It's not that I want to push (well, yes I want :P), but I wouldn't like to see a multi-years development cycle. O:-)
I clicked preview and it posted it...oh well it's formatter properly. lol
Good questions, last time we tossed Beta around was March? Was just something to help keep us focused but seems like it might be a good time to set a new goal.
Agree ... we started what 8 months ago or so ... don't remember for sure, but don't want to be sitting here next year asking if we have a Beta data yet :P
The only option I know of was the Decoda lib ... I have that branch around somewhere, the issue was it was simply slower (50% or so on avg) than whats in there now ... granted it was far more robust and extensible etc ... I was unable to get it to go faster so if someone wants to take a look I can find the branch.
Don't know ... there is always moar there, but where we stop for a 1.0 I don't know.
Ant seemed to feel, as I recall, that the bulk of things were done and it was down to cleanup and full implementation. May need some examples of what needs to be done to the templates or createList things so everyone can pitch in. I don't think anyone wants to start that unless it will be useful (meaning its ready to be done)
Let me try that (it did not post)
Are there other new items we want to add, or is it down to completing what has been added / started to date, there are some gaps and tweaks to things we have done I feel.
Strange...
It may happen with the keyboard (tab-tab-enter), but apparently I can't reproduce it any more...dunno, maybe I was just dreaming. :P
I'm currently working on OpenID 2 implemantation, hopefully I can get it working within the next few days..
Beta sounds good for me, there's still lots of refactoring to do (queries in the controllers, implementing the DataValidator class and so on ..) but IMHO we should push the remaining refactoring to 1.1 and go for beta soon.. Elk shouldn't go vaporware ;)
Cool! :D
I'd like to change so many things... lol (for example admin boards and themes pages), but at least they work for now. :P
I'm a bit fighting between myself on the meaning of "beta", really.
I tend to think that beta may mean: okay, the db and file structure will not change any more (and that will upset Norv :P), then if there is something inside the files to move around...well, we can still play a bit...hopefully without breaking too much. :P
Also, I think version 1 will be far from perfect, but not because we don't care about details, but simply because we will not have an in-depth testing by a critical mass of users, seems normal to me.
Yup, I think it's pretty much at that point (lotsa boring stuff, mostly just chasing down strays). I've had to spend a few days on other things (funny, that) but will get back into Elk stuff tonight and take a good look around.
Probably time to start drawing up a hit list. I know admin and moderation are still basically just 2.1. There's a fair bit of markup stripping that could be productive in there. Should be feasible to do it all while replacing crudy old catbg's n stuff. The basics were pretty sound IIRC. It was just a case of too many tags for not enough benefit.
Ant, do you remember if the separation you were working on (structure VS color) is complete?
No, it isn't. TBH I haven't looked at Elk coding for a few weeks. Needed some time for other stuff and that sorta snowballed. Also, my connection is running slow this week due to exceeding the limit (mobile broadband, no landline out here) so I wont be able to download and upload lotsa megastuff until next week, unless I want it to take hours each time.
I'll revisit the code anyway. The separation was only partially done, but should not be too hard to finish. IIRC, it was mainly some of the styling for generic classes (bars and backgrounds) that still needed tidying up. Main menu, linktree, header, footer, buttonlists and several other areas were already done, IIRC.
The bits that had been split were duplicated over to the bare bones basic variant too. I'd made a start on that. Hadn't had time to do anything significant with a dark variant, but I think it currently has placeholder code copied from the default variant.
PS: I must say that this site runs really fast and clean even with my currently crappy connection. Very nice compared to most sites.
I have always wanted to say this too. It's pretty much the fastest site I browse. Though if I remember correctly smcore forum which was running SMF was very fast too. I wonder if it's because of the server rather than the software?
Yeah could be. Hard to tell without a fair comparison (equal content, load, server, blah blah).
Just to be sure we are on the same line:
development => do whatever we want (add new features, drop features, rewrite large portions of code for whatever reason, change the theme, whatever);
beta => no new "features" (even though the definition of feature is not yet defined :P), "normal" rewrite (i.e. reduce code duplication re-using our functions, for example migration of thins to createList O:-)), try to avoid changes to existing functions (not mandatory, if something can work better in another way better fix it), polish the theme;
release candidate => almost done, only bug fixes unless of obvious problems (if we are lucky it may just be a matter of change the version number... aren't we 8)? ).
final => oh well, at that time we should be already actively developing the next version so, who cares, that's already old! :P
Do you see the release cycle differently?
Thats pretty much how I see it, we could all lots of fluff around that, in general thats kind of it.
A couple of minor notes, I would not call one of the steps development, anything before GA is in a development phase and the steps are generally Alpha, Beta, RC ... then you go to a release phase , GA, Maintain, Sustain, EoL.
Beta should be feature complete, but that does not imply feature working / stable :P ... .Just that you are done adding new things and should be focusing more on performance, stability, bugs. I think for what we are doing, continued MVC fixing, createlist, even extending the newly added features in to all areas (completeness) is all fair game ... but a sliding scale, more of that at the start, and towards the middle and end of the beta cycle it should be all around bugs/performance/stability.
Stickied because I think it need to be starting from now.
Okay, practical things to test:
1) install both mysql and postrgesql (sqlite is now in tools so not mandatory, but nice to have),
2) upgrade/conversion from SMF (I think upgrade is possible and I agree with Norv it would be very, very nice to have, though I wouldn't spend a year on it: if it works fine, otherwise we can go with the conversion),
3) main functionalities: well, it's a long list... generalizing, every button in BoardIndex, MessageIndex, Display and PersonalMessage plus at least 80% of those in the admin panel (yes, for a beta I think a 20% of failures in some obscure admin page is acceptable :P),
There are few automated testing in the tests directory, I just pushed some fixes.
It may be nice to add more tests, but...ehm...I suppose we can do it during the beta phase. O:-)
what is the difference between upgrade and conversion?
Upgrade means basically just "upgrade" the same database with the new informations (so that to move from SMF to Elk would just be "upload the files and run upgrade.php", like upgrade from SMF 1.1 to 2.0).
Conversion is instead a bit "unlinked" because you have do an Elk install and then you "import" the data from an SMF forum.
Okay, today I moved any issue that was not really something broken to 1.0 final.
At the moment there are 7 issues open: 3 are bugs (and likely blocker), 3 are organizational (e.g. where put the news files), and one is a tracking to test the upgrade from SMF.
I think we can add the last one to the known issues of the beta for the moment and work on it later (and even consider to scrap it if not feasible... it would be an hassle though because I couldn't upgrade and I'd have to convert. :P).
Man, does that mean you're all going beta before I do?
Oh my, I just realized I'm the new SMF 2.0 alpha/beta/rc, lol!
Good though, at least I really can tell people to use Elk (rather than SMF) if they think Wedge is too hardcore to them. ;)
Let's see, we may be the last arrived and the first to deliver. O:-)
You're also less ambitious. Well, more than SMF :P
I don't know about your MVC changes though. I'd live to say it makes sense to me, but Wedge dropped pgsql and sqlite so controllers, mehh. Interestingly I'm so used to the folder structure that when I tried removing theme support I was put off by the change. One of the reasons I'm uncomfortable with Elk I suppose. Also the software name (I loved 'Dialogo', too.)
Sorry for the ot. ;)
No problem, at my forum OT is normality, so I'm used to them. (I can't remember a topic that remained in topic for more than 5 posts in the last year LOL)
Well, controllers is not really about the support of multiple databases, more about do the "workflow" and let "others" do the work (i.e. controllers define the "what to do", models the "how to do" if I understood it correctly). The more evident consequence in SMF codebase is that all the queries should be moved somewhere else because queries are obviously something related to the "how to do".
Being 1.0 completed, I guess the targets can be unpinned. :D