[Kstars-devel] Replacing file-system by database in KStars

David Narvaez david.narvaez at computer.org
Wed Jan 29 13:51:46 UTC 2014


On Tue, Jan 28, 2014 at 9:21 PM, Vijay Dhameliya
<vijay.atwork13 at gmail.com> wrote:
> Hi Albert,
>
> Replacing file system by database not only reduce loading time and files
> from source code but it will provide scope to feature like adding, editing,
> and removing DSOs (deep sky objects) and other similar sky-objects.

I haven't used KStars but I don't see how is editing impossible using
the right set of widgets. After all, regardless of whether you load
the data from a file or a DB, you end up with a model and I guess you
can put whatever visualization is needed on top of that to have an
editor.

> And if we have database system in KStars then maintaining user's logs and
> storing downloaded images will become more reliable and systematic.
>
> And KStars allows user to add their own catalog which may contain any number
> of DSOs (i.e. raws for database or line in file). So when talking about such
> large data, loading time will be reduced significantly.

I guess Albert's question is if you have numbes to back that claim.
"Significantly" doesn't have much meaning if you want others to buy in
to your suggested change. One way of getting those numbers is to
implement your changes in a separate branch and then use profilingn
tools that will show exactly how much time is reduced, etc. If you go
down that path I would recommed getting a hold of Milian Wolff who has
done a lot of profiling for KDevelop[0] and other KDE projects. Otoh,
not everything is measured in bytes and seconds, and if these changes
improve code quality, that is another dimension to consider, but
again, that is easier to see once the changes are implemented in their
own branch.

David E. Narvaez

[0] http://milianw.de/blog/katekdevelop-sprint-vienna-2012-take-1


More information about the kde-edu mailing list