open source pouet
category: residue [glöplog]
yeah, tabs could make sense. And yes, linking to somewhere other than youtube preferentially when there is a non-youtube option would be nice.
Also . . .
"Are you hiring Gargaj at a competitive industry rate to deliver new Pouet features on a specific timescale and budget?"
Worth remembering.
I'd like this more if I knew he had a team on it. I wish I was in a position to volunteer because I would.
"Are you hiring Gargaj at a competitive industry rate to deliver new Pouet features on a specific timescale and budget?"
Worth remembering.
I'd like this more if I knew he had a team on it. I wish I was in a position to volunteer because I would.
Metoikos: Oh yeah. I'm definitely not telling anyone what to do here, just responding as the question (on embedded video) was asked, and this is one of my most-visited sites (probably 2nd only to newsblur). So, it's something I care plenty about, and would like to see done well.
First thing has to be a stable code base though. Once that's done fancy stuff like this can be added. Sometimes it helps if there's a good idea of what's needed when cleaning up the code though, because you can leave things in place while doing the re-write.
First thing has to be a stable code base though. Once that's done fancy stuff like this can be added. Sometimes it helps if there's a good idea of what's needed when cleaning up the code though, because you can leave things in place while doing the re-write.
(And I'd certainly volunteer if I had any php/web skills these days - unfortunately the last time I did any web coding was a long, long time back).
Guess it's up to gargaj to decide what happens on that side though, maybe he'll want to single-handedly code pouet 3.0, maybe he'll be so sick of that messy code base by then he'll just want to dump it on git hub and go take a long shower :) Either way, huge respect for taking the job on in the first place. This is a very important site for the scene, and it's essential that it stays at least reasonably up to date, even if that does mean we have to have Facebook buttons ;)
Guess it's up to gargaj to decide what happens on that side though, maybe he'll want to single-handedly code pouet 3.0, maybe he'll be so sick of that messy code base by then he'll just want to dump it on git hub and go take a long shower :) Either way, huge respect for taking the job on in the first place. This is a very important site for the scene, and it's essential that it stays at least reasonably up to date, even if that does mean we have to have Facebook buttons ;)
= 
gloom,
I wasn't talking about duck-taping an untamable beast, but about recreating all outdated or faulty parts without disturbing the user in order to provide a future-proof platform to work on, implementing new features at the same time (as far as possible/sensible before refactoring stuff).
If pouet's codebase really is FUBAR, and I trust Gargaj's judgement on this one for sure, then that would indeed be the most sensible thing to do.
gasman,
My intent was neither to compare the hopefully fun project "pouet dev" to any professional (as in very high-effort and high-quality) or industry/enterprise work nor to tell Gargaj what to do after all, but to point out primarily to others that dropping and recreating stuff people are mostly happy with has far too often damaged the overall project to the point of no return. A lot of thinking and doing and experience has been put into stable code, and that's nothing that, say, two and a half hours of fiddling with Rails can replace. If people are 80 % ok with something, radically changing a few things here and there might drive half of them away because they feel offended after years of being abiding users.
And while I dislike your bitchy tone, your statement is of course right in principle. :)
msqrt,
I second that. But the thing you gain with this approach for that cost (time and effort) is a continuously running platform that does not suffer from lack of time that is spent working on "the thing that will make everything better", but never sees the light of day.
Quote:
Also, professionally, "just fix bux and refactor as long as you can" is a really disasterous approach :)
I wasn't talking about duck-taping an untamable beast, but about recreating all outdated or faulty parts without disturbing the user in order to provide a future-proof platform to work on, implementing new features at the same time (as far as possible/sensible before refactoring stuff).
If pouet's codebase really is FUBAR, and I trust Gargaj's judgement on this one for sure, then that would indeed be the most sensible thing to do.
gasman,
Quote:
I knew I was asking for trouble with that STFU comment... Sorry Gloom, that wasn't directed at you, it was directed at the people offering 'advice' about what Gargaj should be doing with his time.
My intent was neither to compare the hopefully fun project "pouet dev" to any professional (as in very high-effort and high-quality) or industry/enterprise work nor to tell Gargaj what to do after all, but to point out primarily to others that dropping and recreating stuff people are mostly happy with has far too often damaged the overall project to the point of no return. A lot of thinking and doing and experience has been put into stable code, and that's nothing that, say, two and a half hours of fiddling with Rails can replace. If people are 80 % ok with something, radically changing a few things here and there might drive half of them away because they feel offended after years of being abiding users.
And while I dislike your bitchy tone, your statement is of course right in principle. :)
msqrt,
Quote:
I think he could just slowly fix everything, but in the end it's going to be just as much or (probably) even more work than a fresh rewrite.
I second that. But the thing you gain with this approach for that cost (time and effort) is a continuously running platform that does not suffer from lack of time that is spent working on "the thing that will make everything better", but never sees the light of day.
Quote:
I'd like this more if I knew he had a team on it. I wish I was in a position to volunteer because I would.
I guess that's the initial idea of this very topic: Give hooks for people to join in and work on specific parts of pouet.
But, and that seems to be the core response, it needs a somewhat cleaned code base first and a project leader that can coordinate and assign tasks to volunteers (which would likely be Gargaj himself)¹. However, I doubt that this requires the code to be dumped publicly, as it's about working with it, not showing it to everyone.
¹) I'm sorry if this sounds like software industry bollocks talk again, and believe me I don't like that, but I feel it describes it adequately.
Clarification: Gargaj would be the project leader in that context, not also said volunteers, too (though he generally is one himself, of course).
If the aim is to promote running demos in realtime rather than youtube, I'd prefer to look for ways to push that in a positive way, rather than by leaving the video alternative deliberately 'crippled'.
I don't have any killer suggestions for doing that, but as a start, how about a prominent-but-not-annoying banner, displayed to non-logged-in users (possibly with a close button that remembers its state via a cookie), saying something like "Brought to you by Pouet, your friendly destination for the demoscene: audio-visual artworks you can run on your computer"...? (yes, yes, I know. If you can find a better way of describing demos in 20 words, go for it.)
I don't have any killer suggestions for doing that, but as a start, how about a prominent-but-not-annoying banner, displayed to non-logged-in users (possibly with a close button that remembers its state via a cookie), saying something like "Brought to you by Pouet, your friendly destination for the demoscene: audio-visual artworks you can run on your computer"...? (yes, yes, I know. If you can find a better way of describing demos in 20 words, go for it.)
maybe one of those "outgoing link" pages like deviantart or myspace has ("you are now leaving...") could be fun :D
especially if we add this:
especially if we add this:
:D
I like that think Garg just said.
lol Gargaj, that made my day :-)
some thoughts on topic:
- transfer the pouet project to the next generation (coaching, code reviews, ..)
! yea, a rewrite might be in order
- keep the design (or at least keep it as simple as it is now)
- javascript still only "on demand"
- more prod credz/info like on csdb or bitfellas
- keep the moderation as unobtrusive as it is now (gj, btw!)
some thoughts on topic:
- transfer the pouet project to the next generation (coaching, code reviews, ..)
! yea, a rewrite might be in order
- keep the design (or at least keep it as simple as it is now)
- javascript still only "on demand"
- more prod credz/info like on csdb or bitfellas
- keep the moderation as unobtrusive as it is now (gj, btw!)
i still have the psd for this classic somewhere...
Why not split pouet into a well-defined api and a frontend consuming said api? I'm sure a lot of people would love being able to REST their way through the database to do some cool visualizations and it would make posting lolcats infinitely easier.
sagacity: as i said before, a lot of the problem is that the database is complete arse, so the switch to the next version will come with (i guess) several hours of conversion downtime as well.
... or for YouTube, just present a small "play" button somewhere, open a full-window div and populate it with an iframe pointing to http://www.youtube.com/embed/<VIDEO_ID>?hd=1&autoplay=1&fs=1&controls=0 (like this) - that at least forces people to not watch the prod in a small window inside pouet :)
But yeah, good point with getting too close to a video sharing site. Then again, in contrast to said video sharing sites we DO have executable downloads, lots of metadata, an own userbase and an on-site forum that's not just an embedded Facebook discussion. This should do.
But yeah, good point with getting too close to a video sharing site. Then again, in contrast to said video sharing sites we DO have executable downloads, lots of metadata, an own userbase and an on-site forum that's not just an embedded Facebook discussion. This should do.
[LIKE]
sagacity: While that is a nice architecture, it might not help at all with the problem at hand. I'm also afraid that various fringe groups use browsers to access pouet that might not cope that well with JavaScript-centered frontends.
kb: I like metadata.
@all: for crying out loud: Do NOT rely on Javascript because remote code injection/inclusion is inherently unsafe (always was and always will be). any security advisor w/ at least a few brain cells left will tell you that.
@all: for crying out loud: Do NOT rely on Javascript because remote code injection/inclusion is inherently unsafe (always was and always will be). any security advisor w/ at least a few brain cells left will tell you that.
xyz: Yes, let's have a demoscene where nobody ever executes foreign code on their machine. Sounds like a winning concept!
*sigh* you know, there's a big difference between executing *any* code and consciously x'c code from (more or less) trusted sources.
So much talk, nothing to push, pull, commit or merge - meh.
Gargaj, I know the code might be embarrassing, but imagine getting up and bugs are fixed, features are implemented and Rainbows..
As a person who has worked with Spanish 'code', refactored Polish Quälcode and deals with 'code' in FLAs - I can assure you I've seen worse.
So anybody who has the time[0],- set up a Github, Bitbucket or something else we can work with. Favourably something with a featurelist to vote for, so we can, after Pouet 3.0 has the same features as 1.0, implement these. I mean it, all this talk is nice - but what is gained when no line of code is being written?
[0] I can't, it would be job number three,- so let's help Gargaj getting some RL time instead.
Gargaj, I know the code might be embarrassing, but imagine getting up and bugs are fixed, features are implemented and Rainbows..
As a person who has worked with Spanish 'code', refactored Polish Quälcode and deals with 'code' in FLAs - I can assure you I've seen worse.
So anybody who has the time[0],- set up a Github, Bitbucket or something else we can work with. Favourably something with a featurelist to vote for, so we can, after Pouet 3.0 has the same features as 1.0, implement these. I mean it, all this talk is nice - but what is gained when no line of code is being written?
[0] I can't, it would be job number three,- so let's help Gargaj getting some RL time instead.
mog: i doubt that it's that easy. even if people want to help, somebody needs to direct a bit so things move in the right direction and don't get messed up - at least when it comes to implementing new features.
If the plan is to "just clean up" it might work though.
If the plan is to "just clean up" it might work though.
If it's fucked beyond repair we better just rewrite it, as it is - no new features. And yes, someone has to control that, I'd love to see Gargaj having that position.
Clean up never works, as at least one has the idea to introduce bleeding edge feature X - which breaks everything, if that person vanishes. I'd really recommend to code it from scratch, maybe with an REST API in mind - but nothing more than Pouet already is. And why not in Perl, Ruby or anything else some people are confident with?
Clean up never works, as at least one has the idea to introduce bleeding edge feature X - which breaks everything, if that person vanishes. I'd really recommend to code it from scratch, maybe with an REST API in mind - but nothing more than Pouet already is. And why not in Perl, Ruby or anything else some people are confident with?