Gsoc 2014 ideas

[Jef Driesen]

On 2014-02-27 17:15, Dirk Hohndel wrote:
> On Thu, 2014-02-27 at 16:42 +0100, Jef Driesen wrote:
>> Hi,
>> 
>> For the "Ability to Configure Dive Computers" idea, one of the
>> prerequisites is access to a supported dive computer. I think you can
>> easily drop this requirement. The most likely target dive computer for
>> this feature is the Heinrichs-Weikamp OSTC. It's open source, the
>> available settings are fully documented, and modifying them is already
>> supported by libdivecomputer. But reading and writing the OSTC EEPROM
>> (which is where the settings are stored) is also supported by the
>> libdivecomputer simulator. That means you can test your code without
>> needing an OSTC at all. Even firmware updates can be simulated! 
>> Dropping
>> this requirement will probably make it easier to find a student that 
>> can
>> work on this :-)
> 
> Please edit the wiki - you have an account and can do that :-)
> Would you like to mentor this project?

I doubt I would be a good mentor for this project. The majority of the 
work will be about the GUI part, and I don't have any experience with Qt 
at all. But as I already mentioned, I'm happy to assist with the 
libdivecomputer related parts.

>> If the Gsoc ideas are not restricted to subsurface, but can be 
>> extended
>> into libdivecomputer territory, then I may have another suggestion:
>> integrate the subsurface Uemis code in libdivecomputer.
> 
> I seem to remember that you and I talked about this multiple times in
> the past - and came to the conclusion that because of the weird way 
> that
> the Uemis works (with the FAT filesystem that fills up) current
> libdivecomputer architecture made it extremely hard to add that 
> support.
> 
> So we had agreed to do this with the new API.
> 
> But my memory may be wrong.

For me it's also a while since I looked at the Uemis stuff :-)

I agree that many of the Uemis features (e.g. its fully featured 
logbook) will indeed be out of scope for the libdivecomputer project. 
But "downloading" the dive profiles is something that I think can go 
into libdivecomputer.

The planned changes to the libdivecomputer interface will probably not 
improve anything for the Uemis. The core idea will still be about 
downloading dives, just like today.

Anyway, if you think it's not such a great idea, then I'm fine with 
that. The Uemis is (for the reasons you mention) not very high on my 
list. I just ran across the Uemis hack in the subsurface code again, and 
thought that that could be something for GSoC.

Jef