[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: idea... free voting portals...?



On Mon, Dec 06, 2010 at 03:36:00PM -0500, Yaroslav Halchenko wrote:
> hm, is there any web statistics available for those? or even plain
> apache access logs?

Probably those logs are available on Alioth - I never checked.  I simply
assume that if pages are really interesting one or two active readers
are amongst them and would respond somehow.  But I might be wrong.
 
> at least myself I have mentioned it to quite a few people, so I am
> expecting task pages become increasingly popular.

I'm demonstrating these at about 5 talks per year and ask people
explicitely to tell me their opinion about it.
 
> >   2. You have no idea what background / knowledge the voter really
> >      has.  Getting votes for the nicest web page or the most colorful
> >      screenshot does not help.
> well, we do not have webshots (yet) so I see the possible bias by good
> screenshots, but altogether biased estimates are better than no
> estimates at all ;)

Well, I mean prospective packages where *we* do not provide any shot.  I
mean the voting could be influenced by following the homepage link and
finding screenshots there.  I would rather prefer a "I need this program
urgently for my day to day work".  The problem is: You will not learn
about the motivation of the voter (whatever the vote is).
 
> >   3. I consider the "mail voting interface" (=sending mail to the team
> >      list) as good enough to  enable at least one or two active people
> >      "voting" - so as long as I do not see any single mail there seems
> >      to be no reason to make more effrot into this.
> well... since voting is something what I would ask our users to do in
> their spare time, it must be VERY easy and convenient.  Filling out some
> email sounds to me to be of bigger effort for them than just going
> through the form and clicking

True.  But I personally regard this as an advantage.  Those who would
take the time to fire up their MUA and write some text might have *real*
interest and not just enjoy some voting procedure.  I just assume that
people who are not afraid of some effort will be the more valuable
people to work together with.
 
> > That's another thing:  I have more ideas which are easier to implement
> > and at least with the same positive effect to the overall project.
> would you mind sharing the ideas? ;-)

I should probably work out a proper TODO list.  Please ping me again if
this has not happened in the next couple of days (I try to get out the
Debian Med paper until Thursday ...).

The item on top of my todo list is the weekly DEHS mail where I just
sended an example to Debian Med Maintainers[1].  This is close to done
(some smaller hints were given in response which are easy to implement
and some really stupid encoding issue remains for some non-ASCII
maintainer names - those encoding issues consume about 50% coding time
:-((()

Another interesting thing is to finalise the information about
translation of applications in UDD.  There are some data available which
packages contain translations and the code to import this into UDD is
nearly complete.  However the data are incomplete (only gettext is
regarded no other translation methods) and the data are not yet used on
the tasks pages.  I'd regard it an interesting information if we could
say: This application is translated to your language / the following
languages.

Parsing WNPP information into UDD and using this in tasks pages would be
cool and would save a certain amount of work for prospective packages
(nothing is done in this direction).

Package Entropy Tracker (PET) in UDD would be really cool because there
might be a lot of interesting use cases for Blends sentinel.
 
> well -- original intent was not even to implement it (I just dumped the
> idea on you) but just to find some reasonable online service providing
> such voting service so I could refer our users there

No problem with this.  As I said: I might be wrong in my reasoning and
if somebody likes to implement this it would be cool to see which effect
it has.  But I would regard those tasks above more challenging for my
taste.

Kind regards

     Andreas.

[1] http://lists.alioth.debian.org/pipermail/debian-med-packaging/2010-December/008246.html 

-- 
http://fam-tille.de


Reply to: