[Scummvm-devel] What is happening to the ScummVM team?

[Max Horn]

Am 12.02.2009 um 21:59 schrieb Johannes Schickel:

> Max Horn schrieb:
>
[...]
>
>> Yes, there have been very active talks with the FreeSCI guys on  
>> whether (or rather: how and when) to turn FreeSCI into a ScummVM  
>> engine. All of this can be read in <http://www-plan.cs.colorado.edu/reichenb/freesci-logs/ 
>> > and in the mail archives of their mailing list <http://lists.gnu.org/archive/html/freesci-develop/ 
>> >. This contains more or less all the information (at least all  
>> that I know off - I made sure to discuss things only in logged  
>> places most of the time). Granted, this is not very visible for the  
>> ScummVM team, but at least you can still read up on it.
>
> I heard that "you can read up on it in the logs" argument some times  
> now and I am actually not really fine with it :-). Especially since  
> most of it (everything?) happened in the FreeSCI logs, which I  
> usually do not check.

To clarify: I didn't want to imply that it was OK because it was  
logged (I said a bit later that it wasn't OK, or at least I meant to);  
all I wanted to say is: "OK, we didn't inform you, but at least you  
can read up on it there, it wasn't all done in private emails & IRC  
queries").


>> Strictly speaking, the whole merge thing is nothing new, but rather  
>> more or less an extension of the GSoC 2008 project, and Jordi's  
>> well-known (I think?) work on making FreeSCI more "ScummVM  
>> compatible" (the GSoC project built on that).
>
> No it is nothing new, but since Eugene's last mail talks only about  
> merging the GSoC code into FreeSCI, I was a bit surprised that  
> shortly after it seems we now want to merge FreeSCI into ScummVM.  
> Especially then nothing was said about it.

I was surprised by it, too. I only learned about it relatively late,  
too. As I understand it (I wasn't witness o it) apparently we struck a  
nerve with the FreeSCI folks with that email. For example, they  
apparently had support for SCI1 for a long time, and we didn't know  
about it; that showed them that something is not quite right with  
their PR... ;-).

Anyway, this point isn't meant as an excuse for anything either, just  
want to explain the background a bit.


>
>
>> All in all, I think FreeSCI would be a great addition to ScummVM,  
>> and IMNSHO one that fits much better into ScummVM's original spirit  
>> than some of the other more recent engine additions ;-). I  
>> personally see no reason not to add it, as long as the original  
>> developers are OK with it and actively participate, which right now  
>> seems to be the case. Well, if they ever manage to get around  
>> finishing that cleanup they first wanted to make, right now it  
>> seems to be back to sloooow mode with them.... ;-).
>
> Actually I would also require all of them to subscribe to this  
> development list and at least write some introduction.

At least waltervn is already part of the list and introduced himself.  
And syke has been a ScummVM member for many years (before me, I  
think ;). That leaves lars (christoph is not working on it anymore, it  
seems?). I dunno if he is / they are already on the list, but of  
course he/they would join and introduce themselves if/once a merge  
would happen.


> Of course it would be nice when they would actively take part in  
> possible discussions on this list ;-). Apart I am not against a  
> FreeSCI merge myself, as stated above. I am actually quite  
> interested in KQ7 support, but I guess that is nothing for the near  
> future... :-)

With a SCI engine in ScummVM, you and I and other could contribute to  
it, too ;-).


[...]

>> Part of this is because I myself have little spare time (having a  
>> "real" job these days, even though it's "only" at the uni). I get  
>> so many mails behind the scenes that want to be taken off, so many  
>> bugs that pop up, so many porters and engine authors that need to  
>> propped...  I am tired of this. One thing I just have decided to  
>> try to improve the situation. From now one, I will send all the  
>> email I send to prod people or to handle issues that are not  
>> confidential, to scummvm-devel with CC. All. Including those "hey I  
>> sent you a trivial fix for a bug in your port 4 weeks ago, could  
>> you at least ack it?" mails. By doing that with all mails, I don't  
>> have to decide for a mail whether to do it, and maybe this will  
>> help a bit with comm.
>
> I wonder why people do not use this mailing list and/or the bug  
> tracker to ask about bugs / problems, but rather instead seem to  
> mail people personally.

Well, for folks from the press, I can understand it, they prefer to  
address one person if they ask whether they can publish ScummVM on one  
of their CDs. (And yeah, we probably should start to forward all those  
emails, too ;-).

And for some of the other mails... I think some people are just  
uncomfortable talking about certain issues in the public where  
everybody can see their "stupid" questions.

But for most others, well, I really don't know. Anyway, I already  
said, I'll try to forward/CC as much as possible to scummvm-devel from  
now on. And maybe just let others answer, too.


>>> * Are we still one team? Or is it nowadays rather that "the  
>>> ScummVM team" as an umbrella team and various single engine teams?
>>
>> I am not sure. I don't quite know what the difference between the  
>> two is, either... We always have been a loosely grouped bunch.  
>> There are a handful people who care about general stuff in ScummVM.  
>> The majority, though is happy to work on their engine and/or port,  
>> and seems to little care about anything else. Going so far that  
>> over the years, I repeatedly found code where engine authors or  
>> porters would hack around issues in ScummVM instead of talking to  
>> "the team" to figure out together how to solve things properly.  
>> Many porters only pop in (in the sense that I see commits & mails  
>> from then) when a new release is imminent (often that means "the  
>> day before").  So, that seems to make us an "umbrella" team?
>
> I guess what rather makes us an "umbrella" team is that people are  
> only informed about changes when they are affected. May it be  
> because their code is touched or because they showed interest in  
> something before. It is hard to follow the general ScummVM  
> development direction when you do not hear anything about it. Sub  
> team internal communication might be useful and ok for certain  
> aspects, but I think major changes like now supporting SCI should be  
> discussed with all of the team or at least made public before it is  
> done.

While I agree adding SCI should be discussed (and it hasn't been added  
yet, so people can still discuss ;-), isn't this actually *not* an  
example of something which affects people's work? Adding SCI affects  
no other engine, and ports can just not turn it on.... :-).

Anyway, I disagree that the way we communicate is the only thing that  
makes us an umbrella team. ScummVM has grown very big, and difficult  
to grasp in its entirety. While I'd love if everybody was familiar  
with large parts of e.g. ScummVM's infrastructure code, in reality  
most people on our team only have very little spare time, and they  
focus it on small parts of the whole things. All mailing list  
discussions won't change that.


>
>
>> If you meant: "Hey, are we not a team and make all decisions  
>> together, in consensus?" then my answer would be: Nope, and we've  
>> never been that. As I see it, we always have been a meritocracy.  
>> I.e. "he who gets something done has far more to say than the idle  
>> bystanders" ;-).
>
> Actually I do not have major problems with the mentioned way of  
> decision finding. I have rather a problem, that I can not even be an  
> "idle bystander", because I am informed about it.

Understood.


>> In the past, I very often tried to email scummvm-devel about  
>> something, to get feedback and help me decide how to do something.  
>> My experience is that usually this didn't work out at all, or if,  
>> then essentially with a tiny group (usually 1-3) of people actually  
>> talking.
>
> I remember some of those mails. But I think the small feedback is  
> not a reason to stop doing so. Especially when some of the then not  
> discussed changes would have more people interested for whatever  
> reason.

Maybe, maybe not. I once was idealistic and wrote emails to scummvm- 
devel about almost every endeavor I planned. These days, I am rather  
disillusioned when it comes to that. So, while the small feedback  
might not be a reason to stop doing writing emails to scummvm-devels,  
I also ask myself what the reason is to continue writing them?


>> I do think that we should talk about merge decisions etc. earlier  
>> on scummvm-devel -- out of politeness / courtesy, at the very  
>> least. Although I don't think this will actually change the way  
>> decisions are made much, because most people will stay silent, and  
>> we will not always end up with consensus decision either. But at  
>> least everybody can have the feeling of having been informed in due  
>> time.
>
> Maybe it would not change the decisions, but actually I think the  
> politeness / courtesy argument is rather strong here.

Yeah, then again, I don't feel treated very well myself (I am *not*  
talking about your emails, BTW, I think you brought up very important  
points). And frankly, I enjoy coding, bug fixing, improving things,  
refactoring, technical challenges -- I don't enjoy doing tons of  
administration stuff (like upgrading our Wiki & forums) and herding  
and proding a bunch of OSS developers very much, thank you.

So, I think I should, for the benefit of myself and the project,  
reduce my role and get others to work on stuff. For starters, we need:

* a forum admin (or ateam) who also will look into upgrading our phpB2  
(mainly for security reasons, phpBB2 gets no security fixes anymore --  
in fact till very recently we had an old outdated version installed,  
too)
* a wiki adming (or team) who does the same for the wiki: Installing  
security fixes, and stuff
* another mailman list moderator or two wouldn't hurt, volunteers?

There are probably more things, I'll mention them as they pop to my  
mind.


Oh, and it would be nice to have some people who actively drive  
improving ScummVM. Like, not just fixing bugs and adding engines, but  
people who actively work on bigger stuff (and that includes finding  
others to help with it, not just doing it alone), like
* getting 16 bit gfx support in
* revamping the midi/music driver stuff
* finally getting that keymapper / vkeybd stuff into a usable state,  
somehow (maybe requires a rewrite, dunno)
* thinking about GSoC projects (new GSoC is around)
* coming up with new ways to improve ScummVM and working towards that  
goal, actively



Bye,
Max