Clément Stenac | 3 Aug 22:46 2005

Hello !

Dear all,

I just created this mailing-list to ease communication, as the number of
volunteers was more than doubled since the DWN was published.

First of all, once more, thanks to all of you for helping with this
project.

As some of you found out, an IRC channel is active (well, not very
active, but still) for the project: #debian-descriptions  <at> 
irc.freenode.net. Join it if you want...

A few things:

* I've implemented most of era's suggestions in the
  PackagesDescriptionsReviewTool page.
* Johannes asked whether we should adopt the english or american
  spelling. I'm not sure about that, but I guess american is better.
  What do you think ?

The descriptions are updated daily (from the main archive packages
file). If a description is updated while already reviewed, a warning
will appear in the comments.

I'll add quite soon (read: probably not before friday) a bunch of report
pages:

- packages for which reviewers disagree
- packages for which status is "to fix"
- packages which have been updated since their review (You'll have to
(Continue reading)

Kevin Kubasik | 3 Aug 22:54 2005
Picon

Re: Hello !

> * Johannes asked whether we should adopt the english or american
>   spelling. I'm not sure about that, but I guess american is better.
>   What do you think ?
>=20

My vote is going to be American, more Debian developers live in
America, ideally this results in less spelling fixes.

On that note, whats the possibility of creating a collaborative
spellcheck dictionary (for aspell/ispell) that can then be used to aid
in catching spelling mistakes. I know that anything we rewrite should
be fine, but some of the other packages might be easier to miss, a
spellcheck might be worthwhile. It could be implemented in a similar
fashion, and utilizing something like phpspell....

My apologies, I am starting to ramble, just my $0.02

Cheers,
Kevin Kubasik

era eriksson | 17 Aug 10:20 2005
Picon
Picon

Re: Hello !

On Wed, 3 Aug 2005 16:54:37 -0400, "Kevin Kubasik" <qub333@...>
said:
> > * Johannes asked whether we should adopt the english or american
> >   spelling. I'm not sure about that, but I guess american is better.
> >   What do you think ?
> 
> My vote is going to be American, more Debian developers live in
> America, ideally this results in less spelling fixes.

I meant to respond to this when it was originally posted, but I didn't
really know what to say. Now I know.

Rather than entrench an arbitrary bias, I think we should avoid any bias
at all.

In other words, if the existing description has a discernible bias for
any particular variant of English, just preserve that bias. "Fixing",
say, a native NZ English speaker's English into American English is not
going to buy us any friends.

Instead, I would propose the following.

If you propose any changes, make sure your English is consistent with
the English of the package maintainer's, end of story.

If you have to introduce a bias (say, by spelling "color" in any
specific way) just let your own preference rule if there is nothing else
to go by.

Does this make sense?
(Continue reading)

Kevin Kubasik | 17 Aug 14:34 2005
Picon

Re: Hello !


> Rather than entrench an arbitrary bias, I think we should avoid any bias
> at all.

Especially since plenty of the developers that maintain a large host
of packages are not going to be great friends of ours after we submit
hundreds of bugs/patches.

> If you propose any changes, make sure your English is consistent with
> the English of the package maintainer's, end of story.
> 
> If you have to introduce a bias (say, by spelling "color" in any
> specific way) just let your own preference rule if there is nothing else
> to go by.

Fair enough, either way, any executive decision on that front could be
automated, replacing 'color' is not a difficult task, and choosing a
primary dialect for the Debian project is beyond the scope of this
task.

Cheers,
Kevin
era eriksson | 4 Aug 08:45 2005
Picon
Picon

Re: Hello !

On Wed, 3 Aug 2005 22:46:47 +0200, "Cl=E9ment Stenac" <zorglub@...>
said:
> I just created this mailing-list to ease communication, as the number of
> volunteers was more than doubled since the DWN was published.

Great! Do you have a web interface for the mailing list, and/or an
online archive? It would be useful for newcomers, as well as people who
just don't want to keep their own archive (and/or suddenly go over their
mail quota again, ahem :-)

> * I've implemented most of era's suggestions in the
>   PackagesDescriptionsReviewTool page.

Thanks! I've added some new suggestions for you now :-)

> - packages for which status is "to fix"

I went over the news section yesterday, looking at the results of a
finished round of reviews. For the record, the reviewers were Anthony
and yourself. My initial impression was that many of the "not OK"
classifications were rather low priority. If more than half of the
descriptions fail our review, we are much worse off than I thought.

On one hand, we have the caeses where the description author's view of
reasonable base assumptions differs from the reviewer's.

This is probably more marked in the news section than in some others. If
you are reading the news section, you probably know something about
Usenet, NNTP, etc. (at least enough to decide whether you're looking for
an NNTP server or not). As a partial remedy, perhaps there should be a
(Continue reading)

Kevin Kubasik | 4 Aug 18:00 2005
Picon

Re: Hello !

I know this is more of an interface thing, but bear with me...

Perhaps we should just add a dropdown box (as opposed to the radio
buttons). I have found that probably a good half of the packages I
have marked as needing a repair only receive such a notice because of
some flaw with wording, syntax, or grammar. While it would seem
foolish to have a review team reading over every description and
ignore these, most of them are pretty insignificant, and would take
only seconds to fix.

I would propose that we categorize every package that requires a fix.=20
-Grammar, Syntax and the like (includes instances of poor english,
basically anytime that the content would be acceptable if the author
had bothered to proofread)
-Minor Policy Violation (Including a reference to another program
after a reasonable description has been given)
-Major Policy Violation (In direct undisputed conflict with a Debian
Policy, but salvageable)
-Rewrite Required (Hopefully, this will be a rare classification, but
used when its not worth the attempt to protect the authors pride)

I have no particular attachment to these categories, or this system,
its more just what came to mind when I was reading era's post. This
also allows us to integrate more options into each page without
increasing clutter (page is already to wide for my itty-bitty-low-res
laptop).

On 8/4/05, era eriksson <era@...> wrote:
> On Wed, 3 Aug 2005 22:46:47 +0200, "Cl=3DE9ment Stenac" <zorglub@...=
>
(Continue reading)

Clément Stenac | 4 Aug 18:35 2005

Re: Hello !

Hello,

> Great! Do you have a web interface for the mailing list, and/or an
> online archive? It would be useful for newcomers, as well as people who
> just don't want to keep their own archive (and/or suddenly go over their
> mail quota again, ahem :-)
Not atm, do you know one for ecartis ?

> Thanks! I've added some new suggestions for you now :-)
will work on that from tomorrowon

I fully agree with both you and kevin about the need for a
severity/problem category. I like what proposed kevin, I was about to
suggest something like that.

About the long-lasting "grandma" versus "geek" debate, I really think we
must not try to fit every package in a frame. Trust your own judgement,
according to who the target audience is, and how you feel towards the
description

Cheers,

--

-- 
Clément

era eriksson | 5 Aug 07:05 2005
Picon
Picon

Re: Hello !

On Thu, 04 Aug 2005 18:35:47 +0200, "Clément Stenac" <zorglub@...>
said:
> > Great! Do you have a web interface for the mailing list, and/or an
> > online archive? It would be useful for newcomers, as well as people who
> > just don't want to keep their own archive (and/or suddenly go over their
> > mail quota again, ahem :-)
>
> Not atm, do you know one for ecartis ?

Google brings up a few hits, but I wouldn't be competent to make a
recommendation.

The simple thing to do would be to sign up the list for Gmane, though.

http://www.gmane.org/ -- if nobody objects I'll do it on Monday, and
supply my own archive. I don't think I missed any messages while I was
over quota.

> I fully agree with both you and kevin about the need for a
> severity/problem category. I like what proposed kevin, I was about to
> suggest something like that.

Welp, if the end goal is to prioritize the problems, having a
priority/severity field might serve that goal better than a laundry list
of closed categories -- which would probably require a dreadful "misc"
category and/or updates when we discover new categories, but I'm not
opposed to the idea. Maybe we could even have both? Or just set up an
informal mapping for categories to severities.

... Or something like a "keywords" field with some well-defined
(Continue reading)

Kevin Kubasik | 5 Aug 09:16 2005
Picon

Re: Hello !


On 8/5/05, era eriksson <era@...> wrote:
> On Thu, 04 Aug 2005 18:35:47 +0200, "Clément Stenac" <zorglub@...>
> said:
> > > Great! Do you have a web interface for the mailing list, and/or an
> > > online archive? It would be useful for newcomers, as well as people who
> > > just don't want to keep their own archive (and/or suddenly go over their
> > > mail quota again, ahem :-)
> >
> > Not atm, do you know one for ecartis ?
> 
> Google brings up a few hits, but I wouldn't be competent to make a
> recommendation.
> 
> The simple thing to do would be to sign up the list for Gmane, though.
> 
> http://www.gmane.org/ -- if nobody objects I'll do it on Monday, and
> supply my own archive. I don't think I missed any messages while I was
> over quota.

I count 9 (including this one) I can forward whatever is nessicary.
> 
> > I fully agree with both you and kevin about the need for a
> > severity/problem category. I like what proposed kevin, I was about to
> > suggest something like that.
> 
> Welp, if the end goal is to prioritize the problems, having a
> priority/severity field might serve that goal better than a laundry list
> of closed categories -- which would probably require a dreadful "misc"
> category and/or updates when we discover new categories, but I'm not
(Continue reading)


Gmane