The reasons of having and not having the repository

[Ivan Denisov]

A part of members want to make a repository for tracking the changes of "center" BlackBox.

That are the reasons they should not do this?

Who wants to have the Git or Mercurial repository?

[ReneK]

quoted from http://forum.blackboxframework.org/view ... 0&t=13#p54

Basically, you keep on harping technical solutions where the organisational and practical need has not been defined.

Let me give you an example. The goal of the "center", I proposed, is to produce stable versions of the Component Pascal Compiler and those subsystems dealing with the underlying operating system(s) and the IDE. As Blackbox is an OpenSource Community Project, Alpha- and Betatesting is not a concern for the "center", but for the larger community. Based on this (not universally approved) definition, I approach your claim "we need a Version Control System!"

I do not know, that "we" need a version control. Not at all.

Who is "we" in that context?

Does the average user need Version Control? What for? I say, they do not! Not Gitorius, not Github, not any!

If I were an average user not interested in programming or alpha/betatesting the Framework, but only in using it, I need:

1) The latest stable version
2) All official releases (though this is not too important)

This can be done with simple links, no need for version control, since those users can only download and not upload anything.

Probably they'd want a repository for non-framework subsystems, and for this, a version control could be wise, but this is not the primary goal of the "center", and for the time being, Helmut Zinn's CPC works fine (BTW: we should have a link from blackboxframework.org to CPC!)

What about users who want to contribute to Framework development by alpha/beta-testing, but who are not part of the center?

They need the above, plus the "latest version, no matter if stable or not", but not any other unstable versions between the last official release and the last unstable release. They test the framework, report bugs and possibly even send in fixes.

They do not need version control, as we do not want them to automatically upload their versions in a way that their possible changes become automatically part of the Framework, and they for sure do not need earlier unstable versions.


Do the center members need a Version control
I think so. I can imagine doing the work without version control, too, but this would be error-prone and probably would add a level of organizational complexity we probably do not want. On the other hand, if we use version control only for the center members, we have to ask if it is necessary to host it on the same server as blackboxframework.org. I do not see much benefit compared to using sourceforge.net and other sites providing that service. So, if we want this directly on blackboxframework.org is a question of time and money.It's nice to have, but not essential.

I can imagine that this approach is different from your thinking. I may even be dead wrong. But by doing user stories and analysis, we can decide

*) who is the user?
*) what is the necessity?
*) how can we solve the issue?

The first part in any analysis is to get to the right questions. If someone tells you that X is not to be questioned, you should start out questioning exactly X, if you really want to get the job done.

Version control is such an issue. Managers without technical understanding get high on such topics, because they have heard that "version control is state of the art", and therefore one "simply needs it". They do not reason, why it is needed and tend to make false decisions.

As I said, I may be dead wrong with my actual user story, but believe me, the approach is correct, and if the center members do their own stories and then compare it to all the others, we get more intelligence, more knowledge and better solutions.

You seem to think that "version control" is a must. Would you please write a user story for every user group you can think of, as I did above? And then we can compare and get even the wiser, because we both actually worked on the topic. And if we can't answer the questions, because we still do not know the role of the "center", then we should stay the hell away from making decisions that need more analysis.

To which you, Ivan, wrote:

I agree with your story. Terminal user needs zip archive. The VCS needed only for developers who need it. No pushing you or any. But if a group of developers wants to use it, why we should disable that ability? Better it will be the Bitbucket or GitHub repository for not pay for traffic. But Issue tracking that come with GitLab or Redmine and visualisation of changes in synchronised repository can be useful for "center" decisions documentation and "second round" developers.

I think that we need web-site first, wiki and not hurry with our own VCS. But the flexibility of virtual server will allow to make all this stuff later.

So, who else, aside from Ivan, thinks that we need a CVS at this moment in order to further the mission of the "center"?