[haiku-inc] Re: Contract communication part 2

  • From: "Michael Lotz" <mmlr@xxxxxxxx>
  • To: haiku-inc@xxxxxxxxxxxxx
  • Date: Thu, 17 Nov 2011 14:13:34 +0100

Hi

Patrik Gissberg <patrik@xxxxxxxxxxx> wrote:
> With the risk of me nagging; I think it's time for another blog-post 
> about your contracted work.

Sure. I was actually on it, but then wanted to include something about 
my latest leads, so I delayed it a little. Pretty much all of the "main" 
work (apart from casual bug report handling) I've done since the last 
post still belongs to the same topic, so I didn't really see why to 
split it.

> I suggest some info about what you did at BeGeistert, what you did 
> after 
> that and what's coming up. A picture or two would be nice, especially 
> if 
> there are a lot of text as in your previous post. Maybe a screenshot 
> showing something that you are working on, that is yet to come in 
> Haiku? 
> I think that would be appreciated.

I'm really not sure what expectations you have. I've been in the kernel 
for the past two weeks. There's really nothing I can put into a 
screenshot. If anything I could show a screenshot of a running system 
and title it "a running system (notice the lack of a KDL)"...

> In terms of marketing Haiku, I would love if Michael could finish 
> some 
> tasks. It doesn't matter how hard he works and how advanced that work 
> is; the users (and especially the donors) want to see results. Saying 
> "I 
> looked at those advanced bugs that caused xyz memory to not being 
> freed 
> all the time" isn't the same as saying "Wireless networking is now 
> 100% 
> working. Test the network preflet yourself! And here's a nice 
> screenshot 
> to prove it.".

As others have explained already, there is this huge amount of work 
that is going on mostly behind the scenes that really needs to be done 
to make the system everyday usable (i.e. not randomly crashing because 
of some edge case being hit). I totally understand that this isn't 
easily marketable, but that really doesn't change the fact that it is 
needed. I'm in the fortunate situation of being able to take the time 
needed to dig into issues like these, so I'm doing it.

By the way wireless networking is actually 100% working from a 
technical standpoint of view, it just isn't 100% user friendly yet. 
Rest assured that I'll get back to it. But I fear I see the next 
complaint already coming: What's missing is the Key store that can 
handle arbitrary credentials to be stored from various applications 
(including the wpa_supplicant). It is a new API that needs to be thought 
through carefully so that we don't have to break apps in the future by 
amending it over and over. This will take some time to be finished, and 
it will most likely arrive in one single commit (or one single push 
containing a lot of commits) and it will be really hard for someone not 
reading/understanding the code to judge its impact. Also it will not 
result in anything visible other than a tiny "store configuration" 
checkbox in the network join dialog. However in the long run it will be 
usable by other applications and become a feature you naturally expect 
to just be there without thinking about it. Maybe think about it this 
way: http://xkcd.com/974/

> My anticipation for this contract was that Michael's contract would 
> make 
> Haiku take a huge step towards beta. Now it seems as if Michael are 
> working on finetuning Haiku, work that is usually performed during 
> the 
> beta stage. I must say that I am disappointed.

I guess the lower level kernel work I'm doing right now could be pushed 
out to the beta phase. It doesn't change much though, it will still 
have to be done. Why am I working on it right now? Easy: with the coding 
sprint at BeGeistert I could dig into that area of the kernel with 
Ingo, who has a more intimate knowledge there. I'm continuing now 
because my mind is set into that mode and the knowledge is still fresh 
so that I can be more productive. Pushing that work out to a later 
point in time would essentially mean that I'd have to re-enter that 
mindset then, simply wasting time on refreshing knowledge that would 
still be here for free right now. I deemed that an inappropriate use of 
my contract time.

I'm sorry that you are disappointed with how things move forward. 
However I fear your expectations are just not really realistic here. 
Please also keep in mind that strong signals like this aren't exactly 
helping in staying motivated...

Regards
Michael

Other related posts: