pouët.net

Open sourcing pouet.net ?

category: offtopic [glöplog]
Taking it from Chaos' code might provide a "more stable" basis. Although it probably only mimicks the appearance of the site and misses many features under the hood. But anyway.
added on the 2013-05-08 15:14:08 by noname noname
Well, chaos' version does prods, has a system for proper credtis and a cheap CMS for the fron t page but there's lots of stuff missing. Also, as said, one .py file aka totally unmaintainable for a community project. I just mentioned it for reference to convey my actual point: It's not that hard if you know what you do (and Gargaj as well as Chaos aren't really seasoned web devs if I'm not mistaken, and got stuff running nevertheless).
added on the 2013-05-08 15:17:32 by kb_ kb_
@gloom: You, dopamine-man, are too much impulsive. Relax. :]

Just try to imagine analogue being like you... and figure out the possible results.

And I agree with you in a lot of your posts but, mostly, I do not like the way that you explain your point of view. I understand your concern about the "coup d'état" (cause I share that concern too).
added on the 2013-05-08 15:18:23 by ham ham
Quote:
So I have a suggestion. We, the users of pouet (who've had an account more than a week and ideally done more than post on the random image thread ;) vote for somebody to take over the running of the site, and get it through this little 'rough patch'. A bit of actual democracy and real openness.


I vote for Lator as the role of Ronald Reagan and Kb_ as Nancy Reagan
gloom: damn, you really suck at reading. i don't defend how it happened, he acted like a dick. he had his reasons, some more valid then others, others not valid at all. but it did happen. and as a rule i prefer to live in the present.

dipswitch: ultimately it's his site and his prerogative if he wants to open source it or not. and he already made the decision, it's already in place. even though i completely disagree with the methodology that was used, barging in like he owns the place after being MIA for years and taking things to his own hand when gargaj (and others) seemed minimally reluctant to go along with it. despite all that horseshit, plus the lack of sql injection audit before open source launch which is attrociously stupid, i do think it's for the best that it's open sourced.

there have been several things extremly wrong about this whole thread, but right now what matters to me is salvaging the situation we are left with.

and i don't see how folks reacting as if it's possible to put the skeleton back in the closet will do any good. as far as i see it it can go down 2 ways: analogue keeps doing what he claims to be doing, opening the repo and the data to allow everyone to contribute code patches and use the data again. or analogue getting sick of the majority of the users being against him and walk away.

now lets examine both scenarios:
1) you end up with the possibility of multiple people working to improve pouet here and now, and coordinate it into the long awaited v2 with the much desired database overhaul. and i do hopefully hope that in time gargaj et al will return to help finish what they started in v2. it would be silly pride of them not to imho.
2) you have more flamewar, analogue walks away from pouet dev again, leaving no one with access.

have any of you that are advocating for mutiny actually asked gargaj if he's even interested in returning to his overlord position after what just happened? ofcourse i can't speak on his behalf, but my bet is, despite his hurt feelings towards analogue for the stunt, that he's kind of relieved to not be the headmaster anymore, and be able to focus on other things. and on the end view of things he can still work on v2 if he still finds it important, or anyone else can pick up where he left it, since it's also now open sourced.

so when you look at the big picture the end result isn't all that dramatic imho.

you're hurt that analogue stepped all over you, i get that.

and you don't trust that analogue isn't just running an ego power trip, i get that aswell.

to tell you the truth, honestly i'm not that hurt, not only because i also believe open sourcing is the best way forward, but because i been hating pouet for years already after all the crap i had to put up with when doing admin work, i litterally stopped visiting for some periods and when i did come here it usually made me sick and just want to leave. only when i recently started using the site again (about 1 year ago) and reconsidered contributing again, i found myself locked out and everyone still bottlenecking any development on one person, which is just stupid. so yeah, i welcome open source and being able to contribute to it's development again. and despite hating pouet for all the trolls and gits that annoyed me all these years i still want to make it a better place, maybe due to the importance of the database and a vague sense of belonging that i still nurture from all the headaches it gave me back in the day.

and i'm not that worried about analogue the dictator as some of you see him. i only talked with him like 5 times over all these years but i can imediately tell he's not the "my precious" kind of person at all. and if you would read the thread again without the "but you been gone for years" shock, you might even notice his actions make some sense.

tl;dr: he wanted it open and he was afraid gargaj was closing it down, so he took it to his hands to free it. he was the only person who could do it. in his head it makes sense. just sucks that for everyone else, who is used to gargaj calling the shots, it makes none.

regardless of the shock trauma, life goes on, only now we have a github to contribute when we want something fixed.
added on the 2013-05-08 15:31:46 by psenough psenough
Quote:
regardless of the shock trauma, life goes on, only now we have a github to contribute when we want something fixed.


Yep, let's ignore the fact that the former maintainer just got the shaft. I guess the open-ness of the code makes all that worth it.

"Oops - I shot your husband - here's some candy corn! Friends?"
added on the 2013-05-08 15:41:20 by D.Fox D.Fox
Give the source! Give the source!
added on the 2013-05-08 15:43:06 by vibrator vibrator
ok, felipe, i see your point.

but the whole story just made it painfully clear to me not only what a mess pouet is probably code-wise (as a non-coder, i cant judge that, only sense it through several fuckups while adding prods), but also on what a fragile personal layer it is built on. i don't want to contribute to a website where an admin can go apeshit like that. and it's really painful to realise that after twelve years of active contributing, and there were some years where i went to demoparties and the first thing people wanted to talk to me about was pouet, and i had a really hard time to defend pouet e.g. in the amiga scene, where everyone basically just hated it at some point. and now since a few years, pouet is actually a real multi-platform community, not only due to mine and others' outreach efforts, but also thanks to lotek style and havoc and their atari friends, for example... and now everything seems to go down the drain just because an admin went nuts. this is very sad.

i'll concentrate my efforts on demozoo then, which is run by sane and sensible people, and where the opinion of the content contributors counts just as much as the one of web developers. and i think some of the other active pouet content contributors already have made the same decision.
added on the 2013-05-08 15:44:34 by dipswitch dipswitch
Reheated old coffee still tastes shit
added on the 2013-05-08 15:45:31 by maali maali
dipswitch: you're probably regarding that exodus scenario.
added on the 2013-05-08 15:49:57 by psenough psenough
Quote:
and i do hopefully hope that in time gargaj et al will return to help finish what they started in v2. it would be silly pride of them not to imho.


How about not wanting to work in a company where the guy who started the business retains rights to barge in 10 years after he sold it, decide everything you do is wrong because he found religion, throw your whole lego brick idea out the window and demand you all start making figurines of jesus covered in horseshit instead, and when you tell him it's not a good idea he fires your ass on the day?

Beucase I certainly wouldn't want that. Got nothing to do with pride either..
added on the 2013-05-08 16:09:40 by leijaa leijaa
leia: a) your analogy is wrong. b) you're not gargaj. and c) you should read up on the definition of pride.
added on the 2013-05-08 16:24:19 by psenough psenough
Here's my analysis of the different code-bases:

  • Yes, the Pouët.net 0.9.x code-base is a mess
  • Yes, the Pouët.net 2.0 code-base is much cleaner
  • No, the Pouët.net 0.9.x code-base isn't beyond repair
  • No, the Pouët.net 2.0 code-base isn't the second coming of Jesus either

The problem I see with the Pouët 2.0 code-base is that it fixes mostly things that didn't need fixing. It was nice, sure, but these changes comes with a big risk. And without any formal tests, other than "sure, a friend pulled some cranks and kicked some tires, this wors", I would be skeptical to such a big change if I was the maintainer. The Pouët 0.9.x code-base has been live for a long time, it's quite mature. Ugliness is second to that.

At the time of writing, there's been 66 commits in the last 4 days, the vast majority of them within the last 24 hours (58 if I'm not mis-counting). That's a very healthy progress, and one that a lone developer can only dream of maintaining.

The way I see it, the most sensible way forward is something like this:

  1. Patch up critical problems in the old code-base
  2. Refactor the old code-base into a back-end and an API (while hiding the database nastiness, of course)
  3. Gradually migrate the front-end from the new code-base (this is the part that stands out in the code-base) to the back-end/API, one page at the time.
  4. When everything is up and running smoothly, design a new database layout and have the back-end use this instead.


Of course, 2. and 3. would probably be closely tied together, and probably run somewhat in parallel.
added on the 2013-05-08 16:27:58 by kusma kusma
kusma: Stop being so rational!
added on the 2013-05-08 16:34:22 by Tick Tick
Kusma: Spot on. Very much, this.

The question remains, whether Analogue will keep at it in that pace and if not how much everyone else is willing to commit to the project.
added on the 2013-05-08 16:36:38 by tomaes tomaes
Quote:
Analogue reappearing after so long, and offering to open source pouet. That was really great to see - the site has stayed as it is for so long, and a proper API so we can use the data in other places, awesome!

Then we learn the 0.9 code is an ancient mess that needs a rewrite.

But wait.. gargaj is doing that exact rewrite! Except he's been on it for 5 years, which yes, is about 4 years too long, but at least we can see it running and mostly done.

So gargaj asks for a couple of months to finish it, then we can take it from there. I really thought the thread would be over at that point - the site has been running for years without any real changes, there's no imminent threat facing it, so there's no rush. Wait for gargaj to get his shit together, and move on.

And what happens instead? Gargaj loses database access, as does everyone else but analogue it seems. He's basically given the boot.

Then analogue publishes the 0.9 code *without* doing a code review first, there are exploits visible in the code, and gargaj actually submits a bunch of fixes.

So we're left with the guy who's been running the site (generally trouble free) for the last 5+ years and writing pouet 2.0 effectively kicked out, and the guy who's been missing for years taking complete control and asking our help fixing the old pile of crap code.

Well, fuck that. I don't see any point at all in working on the 0.9 code, beyond fixing security issues or obvious bugs as an interim measure. And I'm seriously unhappy that analogue has taken control of the site like this - it's about as far from "open" as it can get.

So I have a suggestion. We, the users of pouet (who've had an account more than a week and ideally done more than post on the random image thread ;) vote for somebody to take over the running of the site, and get it through this little 'rough patch'. A bit of actual democracy and real openness.

What do you all reckon? I'm assuming analogue and gargaj would be 'candidates', anyone else? And will analogue actually listen to and respect the result?

And analogue, before you accuse me of being the 'gargaj fanclub', I've never met the guy, haven't talked to him on IRC or elsewhere. At the start of this thread, he had my respect for his demos and his pouet admin work. You had my respect for setting this site up and running it in the early years. At the end of this thread, gargaj has won about as much respect as you've unfortunately lost.


word.
added on the 2013-05-08 16:37:43 by Defiance Defiance
Kusma, tomaes: It's a bit sad of you to assume that gargaj didn't test his code or didn't think of all the things you mention.

That's exactly the kind of thing I'm afraid when open sourcing code - people picking it apart without knowing all the facts. This, in retrospect costs time for the developers to explain stuff like "yeah, sure I thought of that" which is really frustrating.

Why not giving him the benefit of a doubt? You never second guessed the stability of the old code, did you?
added on the 2013-05-08 16:43:59 by D.Fox D.Fox
D.Fox: I didn't say he didn't test his code. I said there weren't any formal tests; that's something I can see from looking at the code. I'm not even saying that I believe his code is broken - all I'm saying is that I would never ever trust a programmer not to break something without evidence if I maintained a high-enough traffic site like Pouët.

The old has had plenty of real-world testing. But don't get me wrong, I totally think it would be worth it to write some formal tests for it. At least once we have something that makes sense to test.

This is one of the big benefits of gradual changes rather than monolithic rewrites. You can carefully refactor what you have while making sure things works as they should. And if you find that something has regressed, you can bisect it. With a monolithic change, you'll spend lots and lots of time trying to figure out what subtle little thing changed, causing the house of cards to fall down.

I believe we can find a middle-ground where we can benefit both from the battle-tested (but ugly) code-base, while getting some snazzy, clean front-end code from Gargaj's work.
added on the 2013-05-08 17:24:16 by kusma kusma
kusma: from what I know, that's exactly what he did - refactor, test, repeat. It's foolish to think that a seasoned developer would work any other way.

But anyway. I guess we'll have to wait and see what happens next.
added on the 2013-05-08 17:32:42 by D.Fox D.Fox
I'm so glad someone is advocating a gradual changes strategy. In my experience it is the most secure option in most developments. It's really hard to advocate usually because a lot of coders hate it, unfortunately ..
added on the 2013-05-08 17:45:32 by _-_-__ _-_-__
I am a certified test engineer. Is there something like a functional design document or requirement document ? :)

I would be happy to help out and try to make some formal Pouet functional test scripts / checklists with the appropriate testing techniques.
added on the 2013-05-08 17:49:58 by magic magic
D.Fox: What's so difficult to understand about the word "formal"? Please stop putting words in my mouth.
added on the 2013-05-08 17:56:15 by kusma kusma
lol
added on the 2013-05-08 17:57:33 by SiR SiR
_-_-__: Just remember that on the demoscene, "doing the thing you don't hate" will always take precedence over "doing the thing that makes good business sense".


(which is also my catch-all reply to that Joel On Software article from about 10 pages back, btw)
added on the 2013-05-08 18:02:12 by gasman gasman
Yeah D.Fox, put other things in Kusma's mouth.. you know.. long things.. with purple veins!

*wiggles eyebrows*

*blows air horn*

*buttwalks out of thread*
added on the 2013-05-08 18:05:34 by okkie okkie

login