6
$\begingroup$

I posted this question on building Blender and was kindly informed that it's off-topic and I should direct my question to either blenderartists.org or #blendercoders.

Quote:

Reason being, there is often not a definitive answer, issues like this also require a bit of discussion to resolve and then there is also the possibility that the problem is entirely on your end.

As I find this problematic on a systemic level (that is: detrimental to the Blender developer community's productivity at large, aka huge missed opportunity), I want to make an effort to protest this decision.

Concerning #blendercoders: IRC is an ephemeral media - any knowledge it produces is (technically speaking) instantly forgotten. It is also strongly driven by coincidence and does not scale - The bigger the circle gets, the smaller the chances will be that you meet a person on irc that might have recently produced the knowledge you depend on now. (If that person sees your question through all the clutter and lives in your timezone and is not currently having lunch or at the cinema anyway) Should we really promote this? A platform that gets worse as it grows? A platform that produces duplicate questions by design? That enables access of already produced knowledge only through coincidence? Where the reproduction of knowledge means actual manual labor for persons other than the one searching the knowledge? Really? :)

If anything is to come out of this protest of mine here, I would like it to be that we stop actively promoting IRC as a medium for our knowledge production. It's 2014 and I think we can really do better!

So then ... blenderartists.org! This is already more agreeable, but as far as I am concerned still the wrong place to promote for this. Forums are a 90ies thing, and it shows - I just tried to conjure up the search I would have done on BA to find the solution to my above-mentioned build problem: It leads to a bulky form called "Advanced Search", that is so advanced it does not even feature searching for tags, which is not suprising at all, because this forum does not even support tags in the first place. Falling back to a keyword search "build fedora" brings up a staggering amount of threads that have exactly nothing to do with my problem. Later I notice a faultily rendered select box of sorts (overlapping with another), that would let me limit the search to the appropriate subforum somewhat effectively, were it not unusable in that it only shows 2 subforum titles at a time, while being only navigatable with my mousewheel which unfortunately scrolls over more than 2 entries at a time.

And then there's this problem with policy; for all I know a forum is a place to discuss, and that's absolutely fine with me, but why then am I sent to a forum with my dry technical build problems? I don't want to discuss, I don't want a layout that promotes lengthy pieces of literature without any form of rating or discrimination. What I want is harsh moderation, sticking to the point, a concise layout that labels content in terms of its quality and an environment that enforces people to write as much as is needed to solve a problem, but not a single word more than that. And there's a place for that and it's called Stack Exchange!

And now addressing the reasoning to ban build problems from Blender SE quoted above:

Yes, there is often not a definitive answer - which is why SE allows moderators to close and delete questions!

Yes, a build problem often requires discussion to be solved - but to elaborate is not to argument, and as such I don't see the problem with it.

And yes, the problem might well be entirely on someone's end - but how does that justify neglecting the chance to document the solution for a build problem that might affect eg. 1000s of users building on operating system XY under perfectly repeatable circumstances?


I hope this manages to point out why both IRC and Forums are clearly inferior places of knowledge production and conservation for the subject at hand (build problems), and why - although all the reasons against its inclusion do apply - Blender SE would still be the best place for the community to solve these questions in a contemporary and efficient fashion.

Cheers and thanks for taking the time to read and consider this, regardless of the outcome it is much appreciated!

$\endgroup$
  • 1
    $\begingroup$ You might get more useful answers on stack overflow anyway. $\endgroup$ – user3641 May 27 '14 at 10:03
10
$\begingroup$

I would say no, to this.

  • Issues building Blender often end up being specific to a platform, compiler and the version of Blender you're trying to build.
  • Build environments are a moving target, on all platforms.
  • Build errors are very hard to debug remotely, and there is the chance something is broken on the system, its hard for anyone to tell.
  • There is the bf-committers mailing list for developer topics.
  • This is scope creep on blender.stackexchange, building software is unrelated to using Blender (which adds noise and unrelated content for Blender users).

Having had to help others build Blender before, the reasons it can fail are so varied, that I don't think its useful to have here.

Also there is a fairly high instance of very weird/specific problems, sometimes new developers make modifications to their systems which make error messages very strange (while attempting to build).

Eg:

  • A libraries missing so symlink an older library to the new name.
  • Some change is made to Blender in an attempt to make it build, but that fails (further confusing the error message output).
  • The build fails before of a local change to the code.

Considering the time it takes to figure all of this out, and the value of the question to anyone else, I think its not worth to have build questions.

$\endgroup$
  • $\begingroup$ If there's a clear-cut conceptual intention to keep the Blender SE free from actual Blender development, this is all the reason I need as an answer for this - thanks! With all this being told how convoluted build questions can be, I want to point out that just by taking them elsewhere they won't get less complex, and encouraging people to make a habit out of ephemeral problem solving on irc is still detrimental to our development (pun intended), and eventually should be adressed with better approaches than a forum or a mailinglist (thanks for mentioning it though, currently the best bet imo) $\endgroup$ – Simon Repp May 27 '14 at 9:39
  • 3
    $\begingroup$ I realize this may seem harsh, but as a developer you really need to learn how to troubleshoot your own build environment - link errors, header versions, missing includes, compiler compatibility, modifying the build-system, etc... while its highly frustrating at times. I think learning how to do this is not specific to Blender, which is part of the reason I rather not attempt to address this with blender.stackexchange $\endgroup$ – ideasman42 May 27 '14 at 10:03
  • $\begingroup$ If you want to further elaborate on this, I'd be interested in how you'd rather address building problems then, because the part about why not on blender.stackexchange has actually been sufficiently and satisfactorily covered for me! :) That anyone may have to figure out his build problems on his own is just how it is, but I disagree that this means it's pointless to document specific solutions for the whole community to see. Even if a specific solution does not apply to me I can still learn a great deal from the general approach. Denying people, denying us this knowledge - that is harsh. $\endgroup$ – Simon Repp May 27 '14 at 10:38

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .