Codefreek did very little as leader, he has done nothing recently. He will never get this project done. If you want to, I would advise you to take the lead. I believe he offered it to anyone else somewhere in the thread.
Unfortunately I've got too much on my hands already (just been accepted into PHP.net's docbook team and I'm helping other communties as well) although I think I can spare some time for developing. I'm definitely not in a position to lead however.
Perhaps Salathe can boss us around when he gets back? :) At the moment we need to get those things on my list done before we can do anything - and by the current track record it'll take 2 weeks to get them done :P
Here's a basic idea of some file requirements:
Database Connection Abstract
Template Loader + Parser
User Session Helper
User Profile (+edit)
I have to agree with Village Idiot on this, the project has not made any progress nor indeed will it if nothing is done about it. The project leader needs to LEAD.
HE/SHE should be the one that should take in everyone's input and opinions. Then HE/SHE should decide (using other peoples input and opinions to make the decision) or the very least instigate a show of hands on coding standards, CVS/SVN, project management tools etc.
Taking coding styles as an example if I may: We all know everyone has a difference in opinion about coding styles and yes these opinions need to be voiced (and they have, repeatedly) but as long a these different opinions exist, people wont agree and nothing will be passed as a coding standards document that can be followed, its the LEADER's job to find a standard that everyone can at least agree to use. In light of this new found definition to the term 'leader' the person (in my opinion) should be an experienced developer.
When at work, my manager started a discussion about our coding standards, we all had our say BUT it was she that had the final word and made it official, otherwise we would have been still sitting in the office exchanging opinions 10 months on! and you know what? it worked.
Project Leader isn't a shiny name badge, he/she should do as the name suggests and LEAD, or at the very least compile a first draft of a standards document that people can vote on and as far as i can see this has not even come close to being achieved.
I am not having a jibe at anyone, I'm just saying the things I believe need to be said in order for this project to begin properly. If Codefreak (who started off enthusiastically) isn't prepared to take on the full responsibility for guiding others then he should be replaced and that's no disrespect to the guy. There has been countless documents posted and plans drawn up but not one has had any kind of decision made about it or even a show of hands.
If you disagree, than that's your opinion (there's that word again), but someone needs to take the helm even if my understanding of the project leader's job is incorrect or i have missed something.
Something else i don't understand, why are we not using one of the COUNTLESS free project hosting facilities (bug tracking and SVN built in etc) and insisting on somebody to take the time and the expense of buying and installing something we can get for free?
Just to let you know, I started working on a quick tutorial on how to use SVN (on the command line at first), if we decide to use CVS then most of it is the same.
I would prefer SVN, the reasons for this I have posted in this thread already. I personally think it suits this kind of project better than CVS, especially atomic commits and the fact you don't have to specify the type of file when adding (either text or binary).
The tutorial goes over:
Checking out a project for the first time
I should be more or less the same as CVS too so what ever we decide
mysql> SELECT * FROM `users` WHERE `users`.`clue` > 0;
Empty set (0.00 sec)
Looks pretty smart, does it allow you to add a hook to it so that we can have a script to send an email to all devs when someone commits changes?. The sys admin at work did this with ours and its really useful.
I think assembla.com allows you to have a url to execute when a commit is made, i'll look into it.
mysql> SELECT * FROM `users` WHERE `users`.`clue` > 0;
Empty set (0.00 sec)
Beanstalkapp is pretty decent and has a great interface, however you can't specify custom commit hooks. There are a number of predefined ones (such as message basecamp after each commit etc.), but apart from that there's not really a huge amount of flexibility. You're also limited on space and users, which could be a big draw if the project grows.
Assembla is ok, but personly I dislike it.
Springloops is a great svn host, they offer more space and more users than beanstalkapp, however I don't think anyonomous users can checkout the repo.
Github on the otherhand is absoloutely amazing. They use Git (duh :P) which is totally different from svn, but after a little while you get used to it, but in the mean time there are plenty of useful guides, screencasts and of course the official docs
What's wrong with using google code for the subversion repository (among other things)? Using Git would definitely be another option if people are comfortable using it (or willing to learn, but I think the same is to be said for SVN and CVS), but I quite like the centralised nature of SVN over Git (but that's just me).
As for the business about the lack of a project leader, hopefully someone can step up to the plate sooner or later. However, it's hardly an impossibility to get things underway just now without a single person making all of the final decisions. Just someone, anyone, come to a decision (after absorbing what's been said here) about one little aspect of the project and we can tick it off (even if we do come back and review it later). As the saying goes, pull your finger out!
I'll be back home in a wee bit over a week and am not afraid to make some 'final' decisions then. :)
I just read around 300 posts, only to find out NOTHING has been done..
I started up a open source software in c# a month or two ago. We are on v1.0.7. On the team is me, one other person who knows c#, kalle, ross (if he learns c# lol) and a newbie at programming.
We didn't bother writing up standards, or do this or discuss that. I made the first commit, I made the second commit, my friend made the third commit.. and it just kept on going. Now we have a good solid codebase with over 5000 lines of code.
The one important thing is to document. Document why you are using and instead of || or explain what the terniary operator does.. that way, it'll help people skip that part of the code yet understand it the same time.
We had no problems with who's doing what, and why style to do and how we are gonna do our brackets.
So what talkphp should do is just get something started. Even if its just a simple registration script, or even a SQL schema.. just get it started and commit. And then let it grow. Don't worry about "head managers" and "project leaders".. That will come naturally.
You can ask kalle yourself as he submitted a patch for it (very minor but still). I am sure he didnt have trouble looking through the code because the fact is that its well documented.
If I had the same number of developers as you guys, ofcourse there would be a strict guidelines as to how to code, but all this talk about "project leaders" and stuff is kinda stupid imo. Infact thats what "growth" is. You start out small, just a bunch of you doing some random stuff, you grow to a size where it's absolutely mandatory to have someone managing it. Then if grows to a size where you now have groups developing different parts of the programs.
And all this talk about dedicated servers, the vbprojects or w/e, windows vs linux .. but you guys dont even have a project in mind (mind you, I read the other thread and it seems as you guys are doing a gallery).
I use google SVN. I think it would be PERFECT for you guys. Its a small system with many features and you can comment on each other's codes, apply patches and assign bugs.. what more do you need? are you guys developing the next linux distro? You can use talkphp's storage and bandwith to try things out.