I saw this report on PALM app development and its download status



In less than 30 days , they have more than 30 apps in the marketplace.

This makes me think how some ecosystems have a fast development cycle and ramp-up time to get apps delivered while some take a longer time. Which brings me to the subject of Maemo development platform being so hard (from what I read). Its this factor that has put me off from even starting to dabble in Maemo development - and that fact that I need to have a Linux machine to do it.

Earlier for QT and PyQT I remember I could develop either in Linux or on Windows, but with Maemo being so tied down to Scratchbox (and I cant even understand the intricacies of getting it built for the extras repository), this has been one major drawback I believe.

Is this a reason for Maemo apps being much more limited to a slower start and less volume of apps ? What does the community feel ?

I assume the majority of these apps are not free...?

Commercial developers often work full time (on their apps) and have an extra motivation to have them available - the fact they want to get paid.

I strongly believe that the ease of development strongly influences the ultimate success of the platform.

One of the reasons I like considering Java or Vala for development (powerful IDEs, lanaguages, Vala still tied to Scratchbox, though). And, indeed, why I'm considering proposing a "Developing for Maemo with Vala and a real IDE" talk for the summit.

Good marketing helps too.

Tim

I had whimsically asked for a way to develop Maemo apps in my friendly Visual basic environment, since that's what I've been using for around 20 years. I was surprised to find that Mono *almost* gets me there. If Nokia had put sincere, serious effort into such alternative development environments (as they did with S60) who knows where we'd be right now...

Sounds like qt to me

Btw, that amount of app downloads is impressive, given the low number of devices out there, and the low number of apps...

especially when one get ones product to become a fashion topic...

Those figures are not meaningful in any way. Considering there's no SDK yet, those 30 apps were probably written by Palm partners and such with access to pre-release versions and most likely took a lot longer than 30 days to produce.

It would be interesting to see what's available 30 days after the SDK ships, and especially the free/Free apps.

It does have a steep learning curve, mainly due to the large number of technologies/tools one needs to master beyond a basic understanding of C/C++ (autotools, glib, dbus, gtk and all the various gnome libs etc). For better or worse, very few people outside the GNOME community are well versed in all of these. The training material is a reasonable primer, but just introduces the basics and you have to keep referring to upstream documentation.

And no, Qt is not the answer - it will just be another piece of the puzzle eventually.

Well, the target platform is Linux, so nothing unusual there. Similarly you would need a Windows box to write Windows mobile apps and an OSX box for iPhone apps.

Note that Linux will most likely run fine on any machine you already have, and you don't even need to install it (a lot of people are happily using vmware images hosted on other OSs). "Having a Linux machine" isn't the same costly proposition as requiring a Windows machine for Symbian development or an OSX one for iPhone development.

One reason is that so far there hasn't been an implied suggestion from the platform vendor that you are going to get rich by writing Maemo apps (most existing apps are the results of developers scratching their own itches). And that's a good thing IMHO.

Heh, its no visual c++, thats for sure

I doubt my position is typical for much of anyone, but my biggest obstacle was the difficulty in establishing an on-device build environment. Now since all the other platforms have cross-compiling SDKs, this is clearly more of a *n*x user's perspective than a mobile dev's, but as lma mentioned, we've got a platform rather like a typical GNOME desktop, so perhaps this perspective is slightly more broadly represented among prospective developers?

OTOH, I never tried scratchbox, so I'm not saying it's bad; just the expected hassle of getting everything working right kept me from ever digging in and doing it. (I do hope Fremantle will be nicer to on-device work -- there's absolutely nothing about the N8x0s rendering them unsuitable for moderate dev work, and even less about a new OMAP3 powerhouse.)