Active Topics

 


Reply
Thread Tools
Fargus's Avatar
Posts: 1,217 | Thanked: 446 times | Joined on Oct 2009 @ Bedfordshire, UK
#21
Originally Posted by qgil View Post
Proofreading is one of the main reasons to have the wiki as a master and start having drafts perhaps even before an alpha SDK release.

PS: there is a misspelling in your sentence above.
That's the problem with fat fingers and an N900! LOL
 
Fargus's Avatar
Posts: 1,217 | Thanked: 446 times | Joined on Oct 2009 @ Bedfordshire, UK
#22
Originally Posted by qgil View Post
What if the assumption is that a majority of developers looking to that Guide will be new to Maemo, by default will be interested in the Web Runtime and won't need anything from the rest?

What if about a third will go for the QtCreator path?

What if only a tiny minority will really look at the platform SDK based on Scratchbox, Xephyr etc, and the related documentation?
Well to be honest I'd be heading straight for the Qt and native development path because of the type of development that I want to do and don't want overhead.
 
qgil's Avatar
Posts: 3,105 | Thanked: 11,088 times | Joined on Jul 2007 @ Mountain View (CA, USA)
#23
Looking at the "Lego" part at http://wiki.maemo.org/Developer_Guide_table_of_contents it says "Flow Suggestion - Start with 'upstream' and core linux".

I think this summarizes the approach of the current ToC, and perhaps our collision of opinions. Shouldn't be the other way around? First you have the Maemo API (mostly the Qt frontend) and only if you really are interested in more you go down to kernel etc.

The Developer Guide must put the emphasis on what newcomers need to get started. Any developer with a more advanced knowledge will sooner or later find what he is looking for.

This is why I believe it's worth to start with an intro that even a non-developer could find interesting, continue with easy WebRuntime that even someone that is into web development can give a try, and finally get into native development.

The platform description would be out of scope in the Developer Guide, a separate guide.

I'll draft a ToC in the discussion page since I see my schematic proposal here isn't enough to convince.
 
Posts: 119 | Thanked: 412 times | Joined on Aug 2008
#24
Quim, I understand what you mean - don't forget that we covered quite a bit at the weekend - the ToC only represented a small part of the results. I think it would be wrong to take it as anything more than a data point and some general "focus group" opinion.

I am pleased that some of the things we came up with have resonated though - hopefully that'll be useful.

To your particular point: I don't think what ended up summarised in the wiki represents quite what we meant

So I think that was just hasty cut'n'pasting from a badly formatted email

Better would have been:
2. Flow Suggestion
3. Reference: Start with 'upstream' and core linux

ISTR wanting to have "flow suggestions" as a set of targeted flows for devs from different backgrounds or wanting to do different things (we just had app vs platform but adding web in there makes sense)

Then I think we wanted reference material in 2 parts.

First to see what maemo additions there are to standard linux frameworks (as per examples - and frankly, the less Nokia diverge from upstream, the less content there is here). Much of this would refer out to existing reference docs and would only really discuss local modifications. Essentially Nokia are not authoritative for this.

Then on to the meaty stuff: the maemo specific docs where if Nokia doesn't document it then we don't have a clue

This section was expected to be a reference, not really a book/narrative.
 
Posts: 14 | Thanked: 0 times | Joined on Nov 2009 @ Philippines
#25
Just my 2cents, Focus first on getting to core Linux community up and running on developing applications for Maemo and Nokia devices.

This means do a very good job of documenting all low level apis, publish device driver code and clearly highlight highlight how to access the full functionality of the devices.

Basics

* File system details
* Loading operating system
* Battery management
* "Brick" Recovery.... (Assume everyone will brick their system multiple times and make it EASY to recover)
* DBUS

Hardware Specific

* GSM Modem
** Call Control
** Media Control
** SMS Messaging
*** Text
*** Text with Attachments
** SIM manipulation
* Accelerometer
* GPS
* Touch screen
* Wifi
* Bluetooth
* USB
* FM Transmitter

And most importantly, try to consolidate the location of information, provide a real table of contents, and keep the links fresh Finding information on Nokia sites can be maddening.

Good Luck....
 
qgil's Avatar
Posts: 3,105 | Thanked: 11,088 times | Joined on Jul 2007 @ Mountain View (CA, USA)
#26
VinceC, to be honest I think the primary focus for Maemo 6 are precisely the myriad of developers not necessarely familiar with Linux. Then again most of your points should be addressed by the Qt API.

lbt, I understand the current ToC doesn't reflect the Long Weekend discussions. So what is missing? Let's print now the relevant information before it's forgotten.

By the way, what about some good practices and competitors analysis? What are your preferred developer guides and why do you like them? And viceversa.
 
qgil's Avatar
Posts: 3,105 | Thanked: 11,088 times | Joined on Jul 2007 @ Mountain View (CA, USA)
#27
Also, do we all have the same rough understanding about the Nokia Web Runtime? It counts as real application development, not "web development". See the related presentations at http://wiki.maemo.org/Maemo_Summit_2009
 
VDVsx's Avatar
Posts: 1,070 | Thanked: 1,604 times | Joined on Sep 2008 @ Helsinki
#28
Originally Posted by qgil View Post
Also, do we all have the same rough understanding about the Nokia Web Runtime? It counts as real application development, not "web development". See the related presentations at http://wiki.maemo.org/Maemo_Summit_2009
In my opinion Nokia WRT is clearly tailored to web developers, but to produce applications not web sites or web services .
__________________
Valério Valério
www.valeriovalerio.org
 
Posts: 14 | Thanked: 0 times | Joined on Nov 2009 @ Philippines
#29
Thanks for the reply, my real point is start documenting the low layers first, this is the knowledge only Nokia can provide, the community can provide the high level documentation later or for that matter Nokia can.

This is the same mistake Android is making. Reserving functionality for carriers and close friends until a high level api can be provided to the general development community.

Kind Regards,

Vince
 
Posts: 263 | Thanked: 679 times | Joined on Apr 2008 @ Lyon, France
#30
Hi,

Originally Posted by VinceC View Post
Thanks for the reply, my real point is start documenting the low layers first, this is the knowledge only Nokia can provide, the community can provide the high level documentation later or for that matter Nokia can.

This is the same mistake Android is making. Reserving functionality for carriers and close friends until a high level api can be provided to the general development community.
Funny enough, since Maemo is using an entirely free platform (apart from some kernel drivers and small pieces of middleware) the low level APIs are all available already outside maemo.org - Glib & GObject, DBus, the Linux kernel system APIs, libc, Pulseaudio, Upstart, uboot, etc... are all free components, and they're the lowest level APIs you get in Maemo.

The only Nokia-specific APIs are for Hildon (and in the future the Hildon equivalent in Qt), Qt itself in some sense (although there are lots of docs for Qt out there), and all of the hardware-related APIs (telephony, camera, FM radio, accelerometer, ...). Which are quite high level, as APIs go.

Cheers,
Dave.
 
Reply


 
Forum Jump


All times are GMT. The time now is 17:25.