DrupalCon Portland 2013

Cómo hacer más sostenible el desarrollo del core de Drupal

Greg Dunlap  · 


Extracto de la transcripción automática del vídeo realizada por YouTube.

I just sent a really angry tweet which is a great way to get this started I'm Greg Dunlap I'm the initiative lead for the configuration management initiative for Drupal 8 you can find me on the interwebs all over the place as hey rocker I want to talk

about how we can make our core development practices more sustainable and I'm probably going to focus on one thing but if we have other things to talk about feel free two years ago I gave a core conversation at drupalcon Chicago it was the first ever court

was that only two years ago man that's a long two years and it was the first time we had core conversations and it was much less guy standing up in the front as it was a bunch of people arguing with me and it was really fun and really exciting and so while

I'm supposed to encourage people to run up to this mic to ask questions so that they can be recorded for posterity and I understand the importance of that I also only have about 10 or 15 minutes of presentation and then would like to have the recipe discourse

and yelling and throwing things I think that would be really great so you know so feel free to walk up to the mic and feel free to just yell stuff out too and I'll just make sure that everybody's opinions get recorded for for the video later unless

they're really terrible so your last meal right so so here we go court is big court is really big you won't believe how vastly hugely and mind-boggling big core has become not just in terms of code size but in terms of the number of people working

on it you know the the processes and methods that we need to make court work when there are 50 core contributors sitting around a room and Antwerp is a lot different than when there are 1500 core developers in a room in Munich Germany or as there will be this

weekend i believe 3,500 core developers and not even all of them are here and they're not all core developers i'm just talking about people um maybe we can get them later but the numbers are large and they're much larger than they're much larger

than they were five years ago that's sort of the point things have become so large that even Larry can't code a whole subsystem by himself anymore and and that's that's really something so you know we we started doing some things to address

this this scaling problem for the Drupal 8 cycle and one of the things is that we named initiative leads to lead bit projects for Drupal and so there's myself and there's Larry who's leading the web services initiative and even that was too big

for him and so he spawns some of that off and succored Chris Vander water into taking on that part as the Scotch initiative and then gavar has been running the multilingual initiative and John album has been running the mobile initiative and this was sort

of a way to to insert for lack of a better term middle management into Drupal to you know take people to become essentially project managers for big stuff that's important to drupal's future but um even then you know by the time we got to the point

of doing that it was already not enough to insert that one small layer and by the time we started getting into really digging into some of these things we already needed more people than just that this photo is from a code spur that we held in Boston about

a year and three months ago it was februari of last year and all ready to start getting into decisions about how we wanted whisky alone to look we have three initiative leads in that photo plus a gathering of about you know a half a dozen other core developers

and the project leads of two major open source projects because to the right of greece's febian potencia is the leader of the symphony project who flew to from france just to take part in this code sprint that's unbelievable so one of the things that

happen later in the dribble release cycle is that when we wanted to put views in core they didn't have an initiative lead they had an initiative team and that was actually really successful I wish we had all established initiative teams from the beginning

so we have a Damien club and Tim Plunkett and Daniel veiner and xjm shoe so so I think that was really great and I think that was a valuable lesson from the views team that you know even one person can't run a project anymore they had a cross-functional

team of people with mutual respect who could take on responsibilities as people as people's time or interest or wanting to choke each other abdin waned and so that was one of a think one of the big learns of the Drupal 8 cycle which is that teams work

and that we can't have one person leading projects anymore that we're already too big for that we like grew too big for that before I even tried it and so I know that Shannon Vettes who's been helping us a lot with pming for thee for the initiatives

throughout the process is giving a talk in this room tomorrow at three forty-five and she's going to talk more about building teams about projects and I urge everyone to go to that because it's super important and it's something we're really

going to have to think about in the next release cycle so there's that and then another thing that there was a lot of focus on around is that you know as core grows and people grow those people need a way to work on court for more than just their spare

time will allow because as I said before even Larry doesn't have enough spare time to code an entire subsystem anymore and so that means funding development and so there's been a lot of there's been a lot of innovation around funding models in

this release cycle and I don't and it's interesting because there wasn't really a focus on it it's not like there was a big you know thing where everyone said we have to think about funding for the Drupal 8 release cycle it just sort of happened

out of necessity so you know we still had a lot of people who were sort of funded part time by their jobs to work on core or could work on court in between projects and stuff like that but we also had a lot of stuff like Acquia started their large-scale Drupal

project to fund development I made this big funding drive to fund development which was really successful and I hope can serve as a model for other people looking for stuff in the future the there's been a lot more instances of company sponsoring developers

to go places to do things there's been more people getting hired full time or part full-time with the intention of working on core and things like that and I urge everyone to think more about funding because funding contribution is absolutely necessary

for core to continue to grow I've said it I said it people are probably sick of hearing me say this but right now Drupal has all of the business pressures of amateur software project and none of the dedicated resources because if you want to build a configuration

management system and you've got a software project you hire a dev team and they build it and we don't work that way and but we need to figure out a way to work more that way if we want to be able to get more big things done in future versions of Drupal

because in order to plan them and work them out you need dedicated resources with predictable availability and it's and it's impossible to do anything in a timeline or milestones or anything like that if you don't have them so funding contribution

is absolutely necessary for Drupal to grow one of the things I heard a lot when I was fundraising was that people were reluctant to to give funding for something but then we're going to be able to see for two to three years and that's a real problem

because you know while you can make all sorts of ROI cases to people about why the configuration management system is good for their business and why it's going to save the money and make them more competitive if they're if they are working in a business

which has you know somewhat tight returns and and their margins are really thin they don't have a lot of money to be contributing to something they're not going to get back on very quickly and that's that's and I adjusted by focusing on the

kinds of people who had longer development models on people who are more product based in more competitive industries who who can see longer term investments pay off better but that's actually a big problem because a lot of the people places where the

real money is in Drupal right now especially at the big media companies and the services agencies are they they're not really able to invest that much in stuff that they're going to see so far out and so the Hemi started thinking a little bit about

our release cycles and that's sort of where I wanted to talk about some today but in addition to having it be a problem with funding there's there's another really really big problem with our release cycle right now and I think it's very well

I heard I heard a quote from somebody on a call that we had that really shove this point home for me oh that's okay that's the wrong side I already talked about that here's the other quote I think I'm burnt out I'm literally losing sleep

over this problem and I'm too stubborn to care about my own feelings because I know I'm going to have to live with this for the next three years that's a cliff CC Chris Vander water talking on one of our initiative lead calls about a technical

problem that he was having a great deal of trouble convincing other people to get into core and one of the reasons that that I think we have a lot of burnout in the core community and we have a lot of sort of bike shedding and stuff hitting walls in issues

is because of this it's because people know that they're going to have to live with whatever we build here for three to six years that we're not going to be able to iterate on it that we can't just get something in that's good enough and

work with it and see how it grows in the real world we have to get it in and it has to be right because we're going to be stuck and that's a real big problem it's as a community who needs people to be able to you know keep being refreshed on working

with core and feeling like they're making progress and feel like you know their mistakes aren't so final it's it's it's really contributes a lot to the pressure that we have but on the other hand the three the three to four year release

cycle is super important because it provides the long-term stability that enterprise clients want it's been a reason one of the big reasons that Drupal has been that has been adopted so widely because they can install and and and build a site and know

that it's going to be supported for at least five to six years and so just saying we are going to move to a 12-month release cycle and we can break whatever api's we want to in that 12-month release cycle probably isn't going to fly very well it's

really hard for product based companies it's really hard for companies doing large implementations of Drupal it's it's difficult in a lot of ways and I think it would hurt our adoption rate significantly if we did that so I've had some ideas

that I've started to think about about about release cycles and I'm going to talk about one of them here it's not a proposal it's just sort of some stuff that's been floating around in my brain and I encourage any other people to come up

[ ... ]

Nota: se han omitido las otras 5.695 palabras de la transcripción completa para cumplir con las normas de «uso razonable» de YouTube.