More Developers...

[koen]

If you want a refreshing view on development, check http://www.maemo.org and http://nokia770.com/ . It features a sane development platform, testing tools, a live-cd created by users and OpenEmbedded integration before the product even launched.
We'll just have to see how well the 770 sells and how the community evolves, but I for one have faith in it.

[lardman]

I have read this before and didnt find it left me with any confidence that I would have a development environment anytime soon. This takes me back to my original point I can use epoc tools or windows tools and have an application or utility put together in hours or I can spend the same time starting to setup a development area for the Z which needs perl / python / compilers / an intimate knowlage of other bits to get the framework installed before any development can start.

For one thing, the learning curve for EPOC C++ is significantly steeper than for Linux Qtopia C++ IMHO, if you can learn all of that in a couple of hours then well done .

Yes, it takes a while for OE to build its toolchain, (this does get around the issues of GPL and having to provide source for the toolchain on request) but once you've installed a few RPMs and grabbed the BB code and OE meta-data it really is as simple as issuing 'bitbake <some-package>' and all the deps (which will include the toolchain for the first build) will be generated for you automatically - that said, after looking at the links on koen's post I think creating a live CD with a pre-built toolchain and perhaps just a plain pre-built toolchain too would also be a good idea.

You also don't need the intimate knowledge you speak of - you do need to install these apps, but their use is largely hidden from the user.


Si

[koen]

IIRC, the http://www.nslu2-linux.org/ people are planning to have such a live cd, and allready have a frontend for OE, so you just type 'make' and out rolls your flashable firmware image after a few hours and a few gigs of diskspace.
Adapting that for the zaurus (or any other machine) requires: editing the conf and changing 2 parameters (DISTRO and MACHINE) and set another path to the metadata (nslu2 uses symlinks to part of the data, and you'd want everything).

[kopsis]

OK id suscribe to that.  What do we need then to provide / promote a high level language school for the Z?

A CD distro for the PC with tools ?
A HowTo for yxz on Z ?
A library of code stubs / framework apps for yxz on Z ?
What else ?
[div align=\"right\"][a href=\"index.php?act=findpost&pid=85786\"][{POST_SNAPBACK}][/a][/div]

At one point I considered putting together some kind of "Zaurus Python Programming Guide" but I haven't even been able to find the time to finish and release other projects that are much farther along, let alone start a new one.

Perhaps the best way to actually see such a project completed would be for someone who is actually going through the learning process to correspond with experienced developers and document what they learn as they go. Such an approach has two advantages. One is that it's much easier for developers to find time to answer a handful of specific questions every day than to sit down and rediscover and organize their knowledge. The other is that such a guide written by a newbie rather than a multi-year veteran is far more likely to capture the knowledge that other newbies need in a way they're actually likely to understand

[jfv]

I'm not a developer, just a hobbyist. My experience writing programs for the Zaurus and, more to the point, compiling Linux apps for the Zaurus, is that there is very little difference in programming for the Zaurus and programming for Linux.

EXCEPT....

for the GUI part. The problem is not the Zaurus, the problem is Qt. I guess experienced Qt programmers would have no problem moving to the Zaurus, but learning Qt is the hard part. I confess I never quite managed, although I didn't try very hard. Once I got X/Qt running on my Zaurus and could run my Tcl/Tk programs with minimal changes, I lost the motivation to figure out Qt.

Felipe

[koen]

I'm not a developer, just a hobbyist. My experience writing programs for the Zaurus and, more to the point, compiling Linux apps for the Zaurus, is that there is very little difference in programming for the Zaurus and programming for Linux.

EXCEPT....

for the GUI part. The problem is not the Zaurus, the problem is Qt. I guess experienced Qt programmers would have no problem moving to the Zaurus, but learning Qt is the hard part. I confess I never quite managed, although I didn't try very hard. Once I got X/Qt running on my Zaurus and could run my Tcl/Tk programs with minimal changes, I lost the motivation to figure out Qt.

Felipe
[div align=\"right\"][a href=\"index.php?act=findpost&pid=85914\"][{POST_SNAPBACK}][/a][/div]

This points out another problem: not everybody uses QT on his handheld device. And not every handheld is a zaurus.
The 'ease' of development depends on the target platform. The people wanting to develop for the original sharp image will face the toughest challenge. The lucky people are the users of OZ or pdaX

[lpotter]

qtopia.net has been pretty useful; although there are broken links.

hmm.. where are these broken links??

PM me and I wil fix them.

[speculatrix]

Like I said before, if I knew then what I know now, I'd not have gone the download-tools, I would have gone for a "live cd" boot:

http://www.pellicosystems.com/demolinux/zdemolinux/

[raybert]

The general aim of this thread, when I started it, was to suggest that a starting point for people who wanted to develop was set up in order to encourage more people to learn how to develop and hopefully increase the number of developers.

I agree that would be a fine thing.  My main point though was that I think such a thing already exists in the form of the ZUG FAQs and HOWTOs.  I'll admit though that it is not the best organization one could imagine.  If someone has the time to make improvements on that it would certainly be very worthwhile.

One of the reasons why I suggested setting up a "school" was because I believe that being taught the basic principles is better than going out and finding them. Being taught them saves time and hopefully encourages people to use a common starting point and standard.

The things I learnt best and with greater thoroughness were instructor led, teacher led, classes. I achieved more in less time than through self study. I will admit many self studiers are out there, who want to learn something and spend ages finding all they need and learning that way, but the vast majority of people learn through instruction and that was the aim and point of this thread, to encourage some central point of instruction, basic lessons and encouragement for development.

Well, this is a philosophical thing, of course.  I've personally always favored self-study and I wanted to promote that idea here.  I think there are many advantages to it, such as the ability to work at your own speed and the ability to concentrate on the areas that you're most interested in, but probably the main advantage is that it makes one self-sufficient.  I think that is a powerful advantage in a circumstance likes ours where the community just doesn't have the resources to provide much more than it already does (especially given the variety of directions that one could go in; i.e. platforms, languages, etc.).

Someone suggested looking to Linux books and classes and that is probably a great idea for those who are not already very linux/unix knowledgable.

OTOH, great things can happen sometimes when someone has the initiative to push for something new, and I hope that turns out to be the case here.

~ray

[w14]

I've personally always favored self-study and I wanted to promote that idea here. I think there are many advantages to it, such as the ability to work at your own speed and the ability to concentrate on the areas that you're most interested in ...

So how about a combination of the two and setting up an organised mentoring programme.

Something like, beginner posts his current level, and future interests ... experienced developer agrees to help direct their efforts.

We could develop some kind of framework through which each party is happy that they are not wasting each other's time.

Clearly it would only work if people approached it in the right spirit ...

Mike.

[raybert]

And frankly, anyone who isn't already an accomplished developer on at least some platform has no reason whatsoever to expect this to be easy; and no amount of documentation in the world is going to change that.  Those who want to do must invest the time and effort to learn the necessary skills and there's simply no way around that.  Those who are developers had to deal with the same learning curve and often times had *fewer* resources to help them (the late-comers always have an advantage w.r.t. resources for learning).

Well I used to develop in machine code on Z80s but if your saying I can use those skills to deveolp on the Z then Im afraid it dont work.  

I've nevered worked on Z80s (I was more of a 6502 guy) but I'm not convinced that there aren't elements of your experience with them that aren't applicable to development on the Z or on a modern desktop, etc.  There are a set of core principals that apply to all programming environments and I believe that once a person has a solid grasp on those that picking-up a new environment is just a matter of learning the specifics of it.  To put it another way, the concepts are mostly the same, only the names have been changed.

Most folks dont have 20 od years as a professional programmer.  And alot of applications dont need 20 years of experience to code unless the implication is that to develop anything on the Z you require months of free time to get the simplest application running or need 20 years of experience to fully understand and decode the Z compexities.

I mentioned my experience only as an attempt to establish my credentials to speak to this subject, not to imply that one needed that much experience to get anything done on the Z (or on any other platform, for that matter).  I was attempting to address some comments (and I don't recall who said them) that appeared to be complaining that one had to already be a developer to understand the materials that are out there.  My point being that this is in fact true to a large extent and it shouldn't be a surprise to anyone!  Learning to develop is not an easy thing regardless of the platform one chooses to start on and regardless of the resources one has available (more resources can make it easier but it still requires a commitment of time and effort).   Once you have trained yourself, learning future platforms should be much easier.  The point is simply to not have unrealistic expectations.

...  This is one reason that MS has taken off... You done need to be a good programmer to be productive and develop working apps as the IDE is easy to pickup and there are lots of examples to get you going.

I can't help responding to this one also....  

Do you really think this is a good thing??  I, for one, would prefer to use software that's written by competent programmers.  If what you say is true, then this probably accounts for the all the crappy, bloated Windows-ware that abounds.  (In fairness, there is lots of good Windows-ware also.)

I have to agree there are lots of windows apps that suck.  Lots of them professionaly writen too.  But if the application has been developed in a half decent environment with the right tools its more dificult for the uninitiated to make a hash out of the project no.  Im not talking bells and whistles here just a good solid RAD / IDE which dosnt require a phd to get up and running.

There's nothing wrong with good tools.  I was reacting to the part about not needing to be a good programmer to develop working apps.

Bad programmers with good tools can still develop bad code (and probably will).

Good programmers with bad tools can still develop good code (and probably should).

The existence of bad professionally-written Windoze apps (presumably due to the existence of good tools that allow bad programmers to develop "working" apps) is testament to this principal, I suspect.  

~ray

[raybert]

Yet another reason why a "school" for developers would be a good idea, would point people down the right path at the beginning... ;-)

Ooh ooh ooh [raising hand]!  Can *I* get to choose what "the right path" is??  Please?  Please?  Please?



~ray

[raybert]

If someone is ticked off that someone has this attitude (and I'm no programming slouch), sorry. But my time is valuable and I'm not going to waste it hacking away when I can have fun elsewhere and find alternate solutions for my needs.

In case I am the ticked off someone that you're referring to, I just want to clarify a couple of things.  First, "ticked off" is a bit too strong to represent my feelings; I'm not really angry at anything.  Second, I have no issue with your choices regarding whether it is worth it to you to do any development work or your choices of tools, etc.  I think you should use whatever works best for you.  Additionally, doing development work (including getting over the initial learning curve) takes a commitment that not everyone is capable of making.  There's nothing at all wrong with that; everyone's situation is different and everyone has to make choices.  I only get bothered when I hear things that I perceive to be unrealistic.

~ray

[raybert]

I've personally always favored self-study and I wanted to promote that idea here. I think there are many advantages to it, such as the ability to work at your own speed and the ability to concentrate on the areas that you're most interested in ...

So how about a combination of the two and setting up an organised mentoring programme.

Something like, beginner posts his current level, and future interests ... experienced developer agrees to help direct their efforts.

...

Well, that's a good idea.  Unfortunately, making it an "organized program" implies that folks on both sides have to make commitments to it (primarily time commitments); otherwise, it's not really going to be very organized.  My concern is that our mentor candidates simply aren't capable of making a commitment such as that.  As I said earlier, I suspect that most developers can hardly find the time to work on their own projects (I know that's absolutely true about myself).

Having said that though, it's important to add that I believe most developers would be happy to help others as they can.  I'll speak only for myself: I'd be happy to help by answering the occasional question; but my ability to do so will be spotty at best, since my schedule simply won't permit any more than that.  I do already try to help whenever I can.

Maybe a solution (a partial one, at least) is to create a forum just for newbie developers to ask questions that veteran developers can answer as their availability allows.  Think of this as a one-to-many approach which optimizes use of our current resources (as opposed to the one-to-one approach which I think just won't work here due to lack of resources).  The archive from this forum would prove invaluable to later budding developers and it will probably also provide the raw material to build a new developers FAQ.

We do already have the "Everything Development" forum with a bunch of sub-forums for specific topics.  This new forum would probably be exclusively for newbie questions on all development topics and would hopefully encourage newbies to post their questions there.  (I can imagine that some newbie developers might feel intimidated posting in the primary developnent forums.)  It might also encourage newbies to cooperate with each other in their shared learning experience (yikes, that sounds little too touchy-feeling....    

So, what do you think of that?  It might fall a little short of "developer school" but it may be the best we can do with our current resources.  Perhaps we can name the forum "Developer School" (or "Development 101").  

~ray

[lardman]

Well, that's a good idea. Unfortunately, making it an "organized program" implies that folks on both sides have to make commitments to it (primarily time commitments); otherwise, it's not really going to be very organized. My concern is that our mentor candidates simply aren't capable of making a commitment such as that. As I said earlier, I suspect that most developers can hardly find the time to work on their own projects (I know that's absolutely true about myself).

What about something along the lines of the "learn X language in 24 hours" style books - lessons laid out which can be followed and give people a grounding in various aspects of the black arts (I imagine there could be a few different ones - General Linux, X11, Qt, Python, etc.).

Then if people still need help they could ask their peers, and if needs be some of the more experienced developers.

I think this would be a useful resource (in general) and it would probably produce the kind of thing which the thread was looking for with not too much investment of time required (my understanding is that this is the way companies operate to get people up to speed with languages which they don't already know).

The major issue here is coming up with the 'course', however I think this investment of time would be better than individual mentoring in terms of the number of people it could help, etc.

The question is, as ever, is it possible/probable that such a course could be organised/put together?


Si