Menu
▾
▴
Re: [WebERP-developers] Help needed with pettycash workflow
|
From: Dale S. <dal...@sh...> - 2017-04-08 06:46:16
|
I'm just a user but I'd say project direction organic and determined according to developer need. If you need a certain feature, you first propose and discuss it on the mailing list, then develop the feature yourself and offer it back to the project. > On Apr 6, 2017, at 11:50 PM, "ab...@ab..." <ab...@ab...> wrote: > > Hi all, > > Ok guys, I feel like we are going nowhere with this discussion and we could be talking about the same for ever… I can help with any path you want to take, I like the project because os some other reasons and the usage of SVN vs GIT is not determining on if I like the project or not, even when I think might be hurting project and it was one of the reasons I took so long following the project without acting on it. > > About que questions of the workflow change that started the discussion was the change requested by someone or related to this and about the future works. > Do we have a roadmap for the project? > I have been trying to find it and could not, what is the direction we are supposed to go from here. What is the vision for the future? I am looking for something mode detailed than the guidelines and looking to the future. > > Regards > Abel > >> On Apr 6, 2017, at 10:48 PM, Dale Scott <dal...@sh...> wrote: >> >> Hi Gilbert, I agree there are interesting projects on Gihub. Thank you for the list and your time to present them. I added comments on the projects inline below. >> >>> we do not need to chancing any essential thing from the current way things are done to change to git. That is politics and politics is harder to decide than the actual technical actions. >> >> I completely agree, but what I am trying to argue (unsuccessfully it seems) is that successful projects have learned to deal with the politics, and aren't just about the technical. Politics are hard, no question about it. But politics are how a project largely attracts and keeps community members (I refer to people who don't just grab code and leave, but who actively work with other team members to better the project). I work with webERP because of its benefits to a company, not because of the CVS the project uses (although I would _prefer_ Git and Github over the alternatives). >> >> Cheers, >> Dale >> >>> On Apr 6, 2017, at 4:04 AM, "ab...@ab... <mailto:ab...@ab...>" <ab...@ab... <mailto:ab...@ab...>> wrote: >>> >>> Hi Dale, >>> >>> Sample projects on Github: >>> GLPI-project: https://github.com/glpi-project/glpi <https://github.com/glpi-project/glpi> 36 contributors, no dictator. >> >> From the About page on the project website (http://glpi-project.org/spip.php?article50 <http://glpi-project.org/spip.php?article50>), it appears there are five project leaders, who presumably act as a committee. However, two of them work for the IT company "Teclib", while the other three are listed as independent, which I take to mean that Teclib is the true project leader. I appreciate the clarity the project presents. >> >>> Linux Kernel: has a dictator I guess I do not have to talk about him. He es all the starter of Github. They use kind of a network of trust for committing. Infinite number of contributors. >> >> Linus Torvlads created Git as an open-source tool to manage Linux kernel development. GitHub is a completely separate commercial for-profit company that offers hosting for projects who use Git (in addition to providing other features). My understanding of the Linux kernel development process is that it is very structured, new code is submitted via a pull-request to a lieutenant responsible for the related area of functionality who, if they approve, merges it into their sub-repo and sends a pull-request to Linus for approval and merging into the master repo. Very autocratic, and it appears to work well for both Linus and the project (and very different from how the distributions may operate). >> >>> Home-brew (Mac compatible GPL packages): https://github.com/Homebrew/legacy-homebrew <https://github.com/Homebrew/legacy-homebrew> 5621 Contributors >> >> The project has one designated lead and sixteen other designated maintainers, presumably operating as a large committee under the leadership of the lead maintainer. New contributions are submitted via pull request. I thought it interesting this project uses the BSD license and not the GPL. https://github.com/Homebrew/brew <https://github.com/Homebrew/brew> >> >>> Odor (OpenERP): Company dictator company responsible. 594 Contributors. >> >> New code is submitted to Odoo via pull-request, although I believe most code is written by insiders. Do you use Odoo? >> >>> GitHub-Ldap is a wrapper on top of Net::LDAP to make it human friendly. https://github.com/github/github-ldap <https://github.com/github/github-ldap> 13 contributors, 7 branches >> >> It seems from the committer graphs that over half the code was provided by one person, and two more essentially provided the remainder. I was disappointed by their lack of project information. >> >>> As you see is very diverse range of projects. Literally almost everyone uses some sort of git to develop currently. About how democratic, is up to the members of the project want to make it, is not really important on Github. Is good to have a project leader/dictator to use your own words but we do not need to chancing any essential thing from the current way things are done to change to git. That is politics and politics is harder to decide than the actual technical actions. About if better for single developer the good thing is you contribute to projects and you contributions appear on you profile if public project, etc, but they also apear as part of the project and all the members, etc. >>> >>> Lastly check out this link to se the trends for last year: https://rhodecode.com/insights/version-control-systems-2016 <https://rhodecode.com/insights/version-control-systems-2016>Git reached SVN market share at the end of 2012 beginning of 2013 (About 40% each), now git is about 90% of the projects specially open source. SVN is going to die unless they do something SOON. >> >> There are three types of statistics.... ;-). The Rhodecode plots show that Git is clearly preferred by specific social interest groups, but that is not the same as general market share. I'd like to see how many lines of code are being maintained in the various systems for comparison, considering a selection of active well-used projects. Also I suspect these statistics largely reflect open-source projects, which may only be 20%(?) (WAG) of all software development. I believe most software development is closed-source proprietary, and would not be well represented in the results, if at all. The last company I worked for used an internal Subversion repository, and no one outside the company would have known. >> >> >>> The reason are kind of outlined here: https://www.youtube.com/watch?v=4XpnKHJAok8 <https://www.youtube.com/watch?v=4XpnKHJAok8> >> I personally would not reference Linus for impartial unbiassed proof of anything! ;-) >> >> >>> >>> Regards >>> Abel >>> >>> >>> >>>> On Apr 6, 2017, at 12:36 AM, Dale Scott <dal...@sh... <mailto:dal...@sh...>> wrote: >>>> >>>> My problem with setting up a Github repo for a group is that I don't see long-term viability. I believe there is benefit to a sole developer in hosting their own repo on Github for their own customizations, but also believe strongly that each Github repo must be considered a separate project because, if I understand you correctly, there would still be new code that would never be committed upstream, and also there is no "parent" project committed to merging back code from its offspring. While this situation is manageable for one person, especially with changes that are more "customization" than "core", and maybe also for a small number of people who agree 100% on what "best" code is, I just don't see it surviving as the number of developers increases, and especially if they are working to meet their own individual needs or their client's needs, and don't share a collective focus. >>>> >>>> There is no single right governance model, it has to come from the project members and their goals. I will concede that having a governance model isn't required for a project to be successful, but I also believe that it provides clarity and makes it significantly more likely that a group of remote developers with different skills and agendas will come together in a cohesive project. >>>> >>>> Does that explain my thinking any better? >>>> >>>> >>>>> -----Original Message----- >>>>> From: Tim Schofield [mailto:tim...@gm... <mailto:tim...@gm...>] >>>>> Sent: Wednesday, April 5, 2017 7:03 AM >>>>> To: Dale Scott >>>>> Cc: abel; Gilberto Dos Santos Alves >>>>> Subject: Re: [WebERP-developers] Help needed with pettycash workflow >>>>> >>>>> Hi Dale, I think we are talking at cross purposes here. My understanding of >>>>> what Abel was suggesting was that he uses a sync script to get webERP a >>>>> presence on github, not as a separate project - I may be wrong. One way >>>>> synchronisation Subversion SVN to Github is just a two line script, not sure >>>>> about the other direction (how hard, and whether it would be required). I >>>>> think these days very few active projects are on sourceforge alone, and after >>>>> the DevShare fiasco many removed themselves from that platform. At one >>>>> time Sourceforge was the place to go to find open source software projects. I >>>>> do not believe it is most peoples first port of call these days. >>>>> >>>>> Several people in the webERP community are already using a script to pull >>>>> from SVN to Git as a way of syncing a local patch set of changes for things >>>>> that will never be in the main line. >>>>> >>>>> As for what method of governance is best for a FOSS project, that is a huge >>>>> subject, and I am inclined to think there isn't a right or wrong answer. From >>>>> all the projects I watch I would say that so much depends on the characters >>>>> of the people involved rather than the method of governance, but that's just >>>>> my opinion >>>>> >>>>> Tim >>>>> >>>>>> On 5 April 2017 at 04:00, Dale Scott <dal...@sh... <mailto:dal...@sh...>> wrote: >>>>>> Hi Tim and others, >>>>>> >>>>>>> -----Original Message----- >>>>>>> From: Tim Schofield [mailto:tim...@gm... <mailto:tim...@gm...>] >>>>>>> Sent: Tuesday, April 4, 2017 1:56 PM >>>>>>> To: Dale Scott >>>>>>> Cc: abel; Gilberto Dos Santos Alves >>>>>>> Subject: Re: [WebERP-developers] Help needed with pettycash workflow >>>>>>> >>>>>>> Hi Dale, >>>>>>> >>>>>>>> On 4 April 2017 at 18:10, Dale Scott <dal...@sh... <mailto:dal...@sh...>> wrote: >>>>>>>> P.s. Respectfully and IMHO, I think for "this" to be successful, >>>>>>>> the goal from the beginning needs to be "better" than existing >>>>>>>> repos for some very compelling reason, not just its on github with >>>>>>>> easier branching >>>>>>> and merging. >>>>>>>> ;-) and think an auto-synch script would not be a good strategy >>>>>>>> for a project. IMHO >>>>>>> >>>>>>> In my opinion I think you underestimate the real power of Gits >>>>>>> branch/merge functionality. Using Git was first suggested on the >>>>>>> webERP mailing lists years ago when I was the administrator, and the >>>>>>> reason I was given was the branching and merging. I wasn't very >>>>>>> impressed with this as a reason and said so. I was wrong. :) I am not >>>>>>> a huge fan of github though, but I feel it is much better than >>>>>>> sourceforge which I think has been broken for years. >>>>>> >>>>>> I've been supporting Subversion (compared to Git and Github) because I >>>>> don't believe Git is the missing piece to an otherwise successful project. I also >>>>> don't think Sourceforge is "broken", I concede for a lot of work Git is vastly >>>>> more efficient (e.g. branch/merge, distributed work, GitHub pull-request, >>>>> GitHub usability features, etc.), but it doesn’t mean you can't do useful work >>>>> using Subversion on Sourceforge. >>>>>> >>>>>> ... >>>>>> >>>>>>> I think that one of the things Git does is to force a democratisation >>>>>>> of a project. Every developer has their own repository, and can pull >>>>>>> in work from other repositories. This tends to ensure the best code >>>>>>> gets used. But thats just my opinion. >>>>>> >>>>>> I agree in general, but I think democracy in the extreme can be just as >>>>> unproductive as a project with an out-of-touch dictator (Phil, if you ever read >>>>> this I mean metaphorically). Git's not magic, and it's not a silver bullet, and >>>>> while distributed development can be a better model for many projects >>>>> compared to centralized development, the only reason "the best code gets >>>>> used" is because *someone* with commit privileges reviewed the code and >>>>> *decided* what code is better for the sake of the project. Going to the >>>>> extreme, if a dozen developers each think their own code is the best, and >>>>> they all have commit privileges to the master repo, is that really going to >>>>> produce the best code? Personally I don't think so. People with a shared >>>>> vision for the future, a clear understanding of what needs doing, and the >>>>> humility to understand that sometimes their ideas aren't what's best for the >>>>> project, is what produces the best code. The way I see it, Git just helps them >>>>> do it. Tools don't substitute for skill, but they can amplify it. >>>>>> >>>>>> The reason I'm not in favour of an auto-sync script is because I think >>>>> accepting commits shouldn't be taken lightly. Code should be reviewed for fit >>>>> with the intent of newProject, reviewed for quality, side-effects, and >>>>> testability (perhaps even having a blanked rule that no untested code will be >>>>> accepted into the master repo), and then consciously accepted or declined. I >>>>> don't think automatically committing new code from webERP into >>>>> newProject is the best way to build respect and confidence for newProject. >>>>>> >>>>>> So to me, just using a better versioning tool isn't enough, but a better >>>>> version control tool, and a well communicated vision, and a bug tracker, and a >>>>> well-defined development and release process, and a visible test process, >>>>> and..., well, that could be enough. My question to you developers (counting >>>>> myself as a community member and user, but certainly not a developer, at >>>>> least not now) is whether you want a git repo that a few like-minded >>>>> developers can work together in, or if you want to create a world-class open- >>>>> source project and the best damn ERP app of its type. (ok, that last bit was a >>>>> bit over the top, but you have to reach for the sky! ;-)). >>>>>> >>>>>> Cheers! >>>>>> Dale >>>>>> >>>>>> >>>>> >>>>> >>>>> >>>>> -- >>>>> Course View Towers, >>>>> Plot 21 Yusuf Lule Road, >>>>> Kampala >>>>> T +256 (0) 312 314 418 >>>>> M +256 (0) 752 963 325 >>>>> www.weberpafrica.com <http://www.weberpafrica.com/> >>>>> Twitter: @TimSchofield2 >>>>> Blog: http://weberpafrica.blogspot.co.uk/ <http://weberpafrica.blogspot.co.uk/> >>>> >>> > > ------------------------------------------------------------------------------ > Check out the vibrant tech community on one of the world's most > engaging tech sites, Slashdot.org! http://sdm.link/slashdot > _______________________________________________ > Web-erp-developers mailing list > Web...@li... > https://lists.sourceforge.net/lists/listinfo/web-erp-developers |
