Re: Half-year review of Stskeeps's work as distmaster
 

I'll start out by saying I was one of those n810 users who always ran latest Mer. I gave up when it became clear that the release of the n900 put a bullet in Mer's head. I really admire Stskeeps' abilities, and he's probably done more than anyone for the open-ness of Maemo (or Meego or whatever it is this week), but I have a really hard time taking anything he says about software at face value.

Stskeeps, I'll address a few of your statements:

Yeah, especially when you start out tracking it for NITs, then you decide it's easier to port that to the SmartQ than it is to implement a settings panel that didn't accidentally reboot the device. The goalpost was always moving, sure, but the users weren't moving it, and Nokia wasn't moving it. But why not, right? The SmartDevices guys sent you a toy! I know n810 owners who were more excited about Mer than anything Nokia's doing, but their only source of information was the subject lines of emails you generated from IRC.

Hardly. If you had just finished up Mer for the n8x0, you could have saved a lot of time working on the Community SSU, because nobody would have cared -- Mer would have stepped in its place. And the fact that Nokia is switching to Qt doesn't mean GTK is going to be removed from the device. And furthermore, even if they had removed GTK, it's pretty damn trivial to repackage it with QtCurve and ship the damn thing as a support library.



Well, congrats. The many, many people eagerly exploring each testing release of Mer and looking forward to any kind of stable release are really glad you've got a job. Now would you mind updating the damn wiki entry to explain that Mer is dead? Why does anyone even have to ask you to do that? That wiki is the forward face of the Mer project. I knew Mer was dying when nobody could even be bothered to post links to the 0.16testing[0-9] releases, but that wiki proclaims loudly that Mer is new and under active development. It isn't. Please adjust the wiki.

And now I'll respond to some of qgil's confusion:

I agree that Stskeeps is one of the most valuable members of the community, but I'd wager his defensive tone comes from dealing with the people who feel burned by the Mer project.

I remember IRC discussions sounding exactly like this re: Mer. MeeGo is open, and as such I support it, but FOSS projects have a tendency to suffer from horrible scope creep. 3d drivers were a great distraction, but since nobody knew what Fremantle even looked like, it can hardly be argued that Mer needed 3d to survive. Distractions like that are what honestly killed Mer and I wonder what plan you guys have in place to keep it from happening to MeeGo. I'm sure Intel will ensure there's none of this port-to-every-platform wheelspinning; that's not their style.

I want to make it clear that Stskeeps is awesome, and is doing a great job. I fully support him and I'm glad you're funding him. But some of us have asked him once before what we could do, we were told to test Mer and report bugs, and then we got youtube videos of Mer booting on a SmartQ 5 and exhibiting the exact bugs we reported the month before.

I know that Stskeeps' work started as a labor of love. I'm happy that he's now making a living doing what he loves. I know that he's good at it, and will continue to be an irreplaceable asset to the Maemo and MeeGo communities.

But the users matter too. In FOSS, the users matter more than ever, because if they get upset enough they'll fork or leave. The Mer project was a failure in the eyes of the users, because the devs got distracted with port-the-alpha-everywhere syndrome, and because the users' expectations were not managed very well. The PR1.2 thread in the Fremantle forum is a perfect example of what happens when you ignore expectation management.

Watch out for scope creep, guys, and keep an eye on the users. You'll be fine.

Re: Half-year review of Stskeeps's work as distmaster
 

And thank you for testing. I'll be answering honestly and directly. I do have a bit different view on hardware support in a project like this, but bear with me.

I do suffer from gadget lust, that is true. Now, admittedly, my speciality is in the lower parts of an OS, porting and in facilitation. I don't know Qt or GTK enough to code me a UI to get out of a paperbag. Now, people did contribute in this area, but I think either 1) I was too bad at facilitating people getting into these areas or 2) the contributors never really came or pulled their weight. It was maybe a bit of both. I have written about this in one of my blog posts about why reconstructing Maemo was a bad idea in hindsight when it came to doubling as a 'backport' for N8x0.

I wasn't supposed to code entire Mer entirely on my own. Did we as a community pull through? Yes, in many areas. Some, not. Did we learn things about the things
we dived into? A lot. So it goes.

Do I need to point out there has not yet been any indication of Hildon and Maemo GTK in a MeeGo (harmattan, meego) context? As well that any status bar plugins, control panels, etc are tied into the GTK architecture in such a way that we don't see Qt based status bar plugins in Fremantle? A lot of code thatwouldn't be reusable in a qt desktop context.


Will do - was thinking the same thing yesterday. It's not the least bit fun for me Mer has gone the way it did. However, I wasn't refering to my job, I was referring to the skills people learnt and now can engage in conversations with skill and experience and merit for their work within mobile linux.

Now, this is where our perspectives differ. I actually expect a shitload of porting of MeeGo to happen to a lot of different devices. If core platform and hardware support is seperate, each new device brings in more people working on the core platform from different device communities, companies, etc.

It's funny you're bringing up the SmartQ example, as it illustrates my point. it -is- the same system on both. But you're right, bugs weren't fixed - either due to my lack of skill or noone diving into the bug.

This I do agree with.

Thanks for your honest comments. Do you have a irc nick I can relate to?

Re: Half-year review of Stskeeps's work as distmaster
 

May I counter that it might be haughty of you to presume that end users' experiences and opinions count for little to nothing in this conversation? What is the distmaster's point of existing if not to BE the distmaster of a dist (pardon my end-userness but I assumed, in this context about operating systems, we're talking about a distribution).

If he was platform focused, then which platform was he focusing on? I'm not sure I got the impression that he focused on any one in particular. Did you mean the 'maemo' platform or did you mean a particular piece of hardware platform?

Assuming the former, that appears to be all for nothing now--which also blows away your point about 'long-term'. Would have learned as much working with uLinux.

If you mean the latter, then all these different gadgets he was aiming for seem to negate your argument about being focused on the platform--and now with the N900 speak, I'm less convinced that the distmaster shares legacy hardware concerns.. while Nokia continues to pump out a newer kernel/software/fixes for their new stuff... where end users couldn't care less what's happening with the distmaster's work.

Clarify for me, if you would please, why I shouldn't feel exasperated and express it as an end user. Then tell me whether the intent was/is to churn code around to learn.. something (what was learned seems, to me, to be of dubious value in highsight) or whether the intent was/is to offer an alternative distribution that we, end users, can use and contribute to in whatever way we could. Near as I can tell, it never got to a point where anyone outside of kernel coding could contribute anything worth doing yet.

Re: Half-year review of Stskeeps's work as distmaster
 

I'll be in Germany fairly soon, so without mobile internet - been answering from my N900 all this time ;) hence, I'll be absent for a number of hours, so don't treat my lack of replying as not wanting to discuss things ;)

Re: Half-year review of Stskeeps's work as distmaster
 

Never got that impression so far.

I would just like to know that, ultimately, there's a reason why I should bother to hold onto this thing they convinced me to buy on the promises of openness and open-source. I still remember old N800 ads claiming 'future-proof because it's based on open-source'. Tinkerer as I am, I don't dabble in coding to the kernel level but I can still manage to compile and run a CURRENT kernel on my old 486DX4/100MHz in the other room... yet I find it incredibly difficult to make my N800 run anything as current. The promises of Mer raised and then ruined all my hopes.

I digress. When you DO get to reading my criticisms, keep in mind that I'm a highly critical bastard (of others as well as myself) and that I don't compliment much. You deserve a large heaping of praise for letting people vent, point out flaws and taking it with incredible grace. I hope you understand that the person on the other end of the terminal is more annoyed with the way Nokia (and all guilty parties involved.. they know who they are) set these things up to fail than with anything you've done or failed to do. For lack of a better phrase: you asked for it.

Take your time and respond when you can but I await your responses with abated breath.

Re: Half-year review of Stskeeps's work as distmaster
 

First of all, let me state that I'm absolutely ecstatic about the things you have accomplished in the past 6 months, and I can't wait to see the results of the next 6. Well done, and thank you!

To keep this in perspective, most of your Mer work was done as a personal hobby before the distmaster position even existed. So the thread title is a bit misleading, but I get the purpose it serves.

The main problem Mer had IMHO was lack of critical developer mass. I was just as frustrated as everyone else by the slow progress, but let's be realistic here. Making an entire Linux distribution is a huge task and requires far too many skills than one person, or even a small team, can reasonably be expected to have. Despite that, you had the conviction to say "this needs doing", roll up your sleeves and get to work (probably getting pulled in all sorts of different directions that had little or no personal interest in the process) setting an example for the rest of us.

If Mer failed, then it was a failure of the community as a whole. Just from a causality point of view, not casting blame or anything. These things happen.

However even if Mer-the-distribution is now a dead end I don't see Mer-the-project as a failure. If you squint just a little bit you can see all its principles living on in Meego, and I think it's fair to say that it would be very different had Mer not existed to show a better way of doing things.

So, again, THANK YOU and keep up the good work :-)

Re: Half-year review of Stskeeps's work as distmaster
 

The problem, which would be addressed by more focus on Mer^2 as a stop-gap until/if MeeGo is a viable day-to-day OS, is that it's now obvious that Mer could never have succeeded in delivering a day-to-day OS for most people.

The apps are the problem. You can see this on the Joggler with Mer 0.17 & OpenGL: you get a Hildon Desktop, you can install some apps, but if it's not ABI compatible with Maemo you're going to have a lot smaller pool of apps. Even if it did get large scale adoption, there're also API issues.

The first few iterations of MeeGo, and the core FLOSS apps it comes with, will give us an indication of whether MeeGo as a day-to-day OS for the N8x0 is viable.

Perhaps Mer^2 and using that as a place for further playing with the GL drivers will help with the MeeGo-on-N8x0 adaption layer. However, this is one that Stskeeps can best comment on - and whether he's getting/got enough assistance from the community on the N8x0 MeeGo h/w adaption layer.

Re: Half-year review of Stskeeps's work as distmaster
 

Someone implied that some criticism has been less than fully informed. True: a general appeal for comments was made, and they are forthcoming. If only well informed criticism was wanted, private messages would have been sent to those who knew more about the actual conditions you worked under.

The point was made that the community wasn't very involved in the development of Mer and thus shares some responsibility for its failure.

Probably so, but I don't think that a good job of momentum-building was done. I remember personally expressing skepticism at whether it was necessary to stop everything and wait for the two drivers. There were other delays that seemed mysterious to me. Once it became clear how great Mer's potential was, there must have been greater and greater willingness to help out.

I think there must have been an inherent divergence between Nokia's focus and the interests of those who wanted to keep their tablets happy. I don't see why Nokia would cry when tablet owners were essentially smothered by those with other preoccupations about the N900 or MeeGo. The renaming of ITT was not just some isolated decison and neither was the foundering of Mer. They were all part of a reallignment of power in favor of the one who was paying the bills.

So the interests of this site were realligned and that made any cries for help you may have made, or requests for discussions of what parts of Mer to continue and what to abandon hard to hear over the N900 and MeeGo hubbub.

That was not your fault; maybe it was inevitable as soon as the true interests of those with internet tablets acquired second-class status.

Re: Half-year review of Stskeeps's work as distmaster
 

The tone is unfortunate. In my interactions with stskeeps, I have seen that he is a tremendous asset. I think he is lightning rod for dissatisfactions that mostly properly lie elsewhere.

Yes, I have pondered this. Mer did not need to be a finished product to influence direction of SW development. Whether in fact this was the case, I do not know.

But there you go moving the goalposts, again, I notice. There is a lot of dissatisfaction in the OS2008 community, and you speak past it. Which of the OS2008 components have been opened? And how will OS2008 community benefit from MeeGo? I don't see it. Will you point to MeeGo adaptation project for N8x0? It is unfinished and will be for quite some time - doubts abound.

I concur. Don't let the crosswinds of criticism deter you.

Thank you for funding Carsten. He is a valuable asset to our community. Please leverage your investment by helping him help the Maemo (not MeeGo) community.

Re: Half-year review of Stskeeps's work as distmaster
 

Forgive me, but this is from one perspective entirely incorrect, and from another perspective slightly incorrect.

From the perspective of "I want my device to continue to run programs at all" you're entirely wrong. Half the time on my n810 I'm running a regular x11 app over ssh forwarding, and I haven't got many problems. I'd be perfectly happy running regular desktop applications on my tablet, if it came down to it. Mer was a great platform for that.

From the perspective of "I want my device to continue to run programs that are perfectly hildonized and integrated" then you're only slightly wrong ...and only because it's possible to fake it a la sugarization.