GSoC comparison

[David Kastrup]

Wow.  I looked at
<URL:http:http://worldofgnome.org/29-new-things-this-google-summer-of-code-will-bring-to-gnome/>:
29 GSoC projects?

Point 1:

    1.) Replace lex-bison based parser with handwritten parser in gcalctool
    The aim of this application is to change current lex-bison based
    parser with the handwritten parser. As, handwritten parser are much
    faster and can be ported to other languages without much trouble,
    this will help in both, making gcalctool fast as well as portable.

"As, handwritten parser are much faster".  Than an LALR(1) table-based
automaton with an O(1) algorithm.  For use in a desktop calculator.  I
have the suspicion that the student will learn more than the project.

    4.) Completion of the Gnome Sudoku Vala Port

    The Gnome Sudoku game, currently written in python, is not up to the
    standards set out. My project will complete the Vala port, and make
    sure its easy and worth to play.

Well, we have our own porting-idea-of-the-month contest going on in the
user list.  I don't even say that porting does not help an application
_if_ the port is done by people experienced with the original code base.
It is an opportunity for redesigning in the light of experience.  And
one has better motivation for _finishing_ an overhaul when one is
porting from X to Y rather than porting from X to X.  I've been in a
commercial project where the build system was ported from Make to Maven
in order to make things work well.  One could have improved the results
for the workgroup further by taking the results and porting them to
Make.  But of course, there would have been no resources or justifiable
incentive for porting from Make to Make.

    6.) GNOME Clock

    Work with the GNOME Design Team and Seif Lotfy on GNOME Clock, in
    order to provide the GNOME community with a clock application. GNOME
    Clock will allow users to quickly and easily determine the time of
    day anywhere in the world, helping them keep in touch with family,
    friends and co-workers. GNOME Clock will also allow users to set
    alarms so that when time X rolls around they know they need to do
    Y. A stopwatch/timer will also be included so that users can time
    themselves and/or others as needed.

Now it's not as bad as the first look: certainly more than half of the
projects are not of this "I could pull out my hairs" variety.  And those
projects were likely accepted under the general GNOME umbrella rather
than individually, so they don't really have more elated status than our
GSoC pitch.  But it does leave me with some head-scratching.

-- 
David Kastrup