I don't think the impossibility of perfection is an argument against better accuracy.
Birgitte SB
--- On Thu, 10/15/09, Cecil cecilatwp@gmail.com wrote:
From: Cecil cecilatwp@gmail.com Subject: Re: [Wikisource-l] Pages without text To: "discussion list for Wikisource, the free library" wikisource-l@lists.wikimedia.org Date: Thursday, October 15, 2009, 1:16 AM Birgitte SB: please read ThomasV's statement a little bit up this thread, the one where he explains why and for what level 0 was introduced in the beginning. That is one of the problems. Each community considers something different to fall among that state.
With the last PR2-update it suddenly was no more possible to switch already finished projects to PR2, because it was disabled for a person to immediately select 'finished'-state. As an example (I did that switch to PR2 just a few days before): http://de.wikisource.org/wiki/Seite:De_Neue_Thalia_Band1_385.jpg was immediately set to finished because the poem was already proofreaded twice (http://de.wikisource.org/w/index.php?title=Das_Bildni%C3%9F&action=histo...). But it belonged to a large story collection (http://de.wikisource.org/wiki/Index:Neue_Thalia,_Erster_Band) which also is just a part of seven such books. They are part of our currently most important community project which should be finished until November 10th. Because of the short time left all books were switched to PR2 so that more people can work on them and also have a better overview which poem/story/... in which book is already finished and which not. I finished the whole switch just a few days before the mediawiki-update. A few days later I would not have been able anymore to set the pages to 'finished'.
When I complained about that to ThomasV, he suggested to use level 0 for it, since it is a finished page and no more proofreading needs to be done. He would even rename the 'without text' to fit it better and the pages still would land in category 'finished'.
Does this sound to you like the statistics of this level would be useable to determine how many pages a project has finished and how many there are that are just empty? Not to me. So those people who for some reason need statistics wouldn't be able to get a real result out of it anyway, because one projects uses it for really empty pages, the other one also for those with less than ~10 words and the next one for pages that do not need proofread for some or the other reason, even if they are full of text.
Cecil
(The other solution suggested by ThomasV would have been to create a list of all pages and which state they are and then let a bot or a second user correct the state of those pages that I would not be able to do. The 7 Thalia-bands have 3279 pages full of text and a few more with index, introduction, headers and first pages, so you can imagine how difficult it would have been to maintain those lists without any chaos and always obtain a format that the bot can deal with.)
2009/10/14 Michael Jörgens joergens.mic@googlemail.com
Birgitte SB that was the reason, why my text wasn't as polite as it should be. It sounded off-topic for me.
joergens.mi 2009/10/14 Birgitte SB birgitte_sb@yahoo.com
Thank you for explaining, this discussion would have been off-topic in the other thread. I am not someone that finds statistics to be a big deal, but I have learned that others feel deeply about such issues. The statistics total all pages marked "4" and report this as the number of pages a subdomain has proofread and validated. So when compared to other subdomains which mark empty pages with "0" the numbers of de.WS are slightly inflated. The current state of things is that other subdomains do not count empty pages in their totals while de.WS does. I appreciate your introducing the situation to the local community.
Birgitte SB
--- On Wed, 10/14/09, Michael Jörgens joergens.mic@googlemail.com wrote:
From: Michael Jörgens joergens.mic@googlemail.com
Subject: Re: [Wikisource-l] Pages without text
To: "discussion list for Wikisource, the free
library" wikisource-l@lists.wikimedia.org
Date: Wednesday, October 14, 2009, 2:29 PM
Sorry, that was my fault, I
didn't read
the topic right, i thougt this discussion is still
part of
the topic [Wikisource-l]
Proofreading.
From
my personal point of view, there is no difference
between
this pages. I see no need to distinguish empty und
full
pages. The final state means, all work needed is done
with
this page.
In
general I like statistiks, but never mind, I would
never
count empty pages.
On
the other hand if others like to do, no problem in
general.
I
think we will introduce this question to our community
and
we will make a decision. Changes can be done by bot,
as I
understand.
sincerly
joergens.mi
2009/10/14 Birgitte SB birgitte_sb@yahoo.com
I sorry you find the concerns other people have to be
"empty" and "silly". Perhaps in
the
future you could ignore the threads on smaller issues
without comment. It would really help keep things
more
congenial.
Birgitte SB
--- On Wed, 10/14/09, Michael Jörgens joergens.mic@googlemail.com
wrote:
From: Michael Jörgens joergens.mic@googlemail.com
Subject: Re: [Wikisource-l] Pages
without text
To: "discussion list for Wikisource, the
free
library" wikisource-l@lists.wikimedia.org
Date: Wednesday, October 14, 2009, 12:15 PM
Is there a need for
this
''empty discussion'' about empty
pages
here,
The main point ist to find an for all acceptable
way of working together with Thomas. In a way
that Thomas is satisfied and feels comfortable,
taking the
problems he see's into concern and the
problems
some
projects have.
And it should be a compromise that all can
accept,
without feeling overruled by others.
With this in the background the other discussion
is simply silly. Some are talking about some
pages,
which
in total are below 1%, there are bigger issues
to
solve.
sincerly joergens.mi
2009/10/14 Birgitte SB birgitte_sb@yahoo.com
I
disagree in general with the idea that everyone
must
conform. There is room for the communities to
develop
their own solutions. Hopefully when one
community
proves
that some solution has good results others will
learn
from
their experience and adopt it. But there needs
to be
room
for experimentation.
In this specific case, it seems to me this is
more of
legacy issue on de.WS rather than a deliberate
choice.
If
they (as well as every other subdomains) conform
to
the
standards the majority of Wikisources use then
the
comparative numbers between Wikisources will be
more
accurate. I don't see a strong reason for
de.WS
deciding to not conform here. But it is a bunch
of
tedious
work and it should be thoroughly discussed. It
would
be
very nice if de.WS took the time to consider the
issue
and
let us know what the consensus turns out to be.
However we can always make a footnote when
people
choose to
count things differently and estimate with data
crunching.
(i.e. Of 1000 validated pages on the top three
non-de.WS
subdomains the average number of empty pages is
X
(A%),
therefore an estimate of de.WS validated
non-empty
page is
the given total minus the A% of the given total)
This is
not an unsolvable issue and I am sure sooner of
later
some
subdomain will discover a good reason to do
something
differently.
Birgitte SB
--- On Wed, 10/14/09, Syagrius syagrius@gmx.fr
wrote:
From: Syagrius syagrius@gmx.fr
Subject: Re: [Wikisource-l] Pages without
text
To: "discussion list for Wikisource,
the
free
library" wikisource-l@lists.wikimedia.org
Date: Wednesday, October 14, 2009, 10:47 AM
#yiv579852079 html, #yiv579852079
{}#yiv579852079 html
{width:100%;height:100%;margin:0px;padding:0px;}#yiv579852079
{font-size:100.01%;font-family:Verdana,
Geneva,
Arial,
Helvetica,
sans-serif;background-color:transparent;background-image:none;margin:0px;padding:5px;}#yiv579852079
p {margin:0px;padding:0px;}#yiv579852079
{font-size:12px;font-family:Verdana,
Geneva,
Arial,
Helvetica,
sans-serif;background-color:#FFFFFF;}#yiv579852079 p
{margin:0;padding:0;}#yiv579852079
blockquote
{padding-left:5px;margin-left:5px;margin-bottom:0px;margin-top:0px;}#yiv579852079
blockquote.quote {border-left:1px solid
#CCC;padding-left:5px;margin-left:5px;} If
the
Wikisources have adopted
some common rules, every sublanguage
should follow it. On
the other hand, I don't understand why
you
don't
want to transform these empty pages, since
en.ws
and fr.ws
already did it and a bot could do it very
easily.
There
would be absolutely no loss of quality or
credibility...
Syagrius
----- Message
d'origine -----De : John
VandenbergEnvoyés : 14.10.09 02:03À :
discussion list for Wikisource,
the
free
libraryObjet : Re: [Wikisource-l] Pages
without
text On Wed, Oct 14, 2009 at 10:12
AM, Cecil
wrote:
Syagrius, could you please explain why
this
would
be
"irrespectuous toward
other wikisources" when we mark
them as
part
of a
'finished' project?
It is not a part of the same work.
Advertisements are a _different_ work, and
it has
not
been
transcribed.
It should be marked as a incomplete.
Advertisements are also sources....
:-)
--
John Vandenberg
Wikisource-l mailing list
Wikisource-l@lists.wikimedia.org
-----Inline Attachment Follows-----
Wikisource-l mailing list
Wikisource-l@lists.wikimedia.org
Wikisource-l mailing list
Wikisource-l@lists.wikimedia.org
-----Inline Attachment Follows-----
Wikisource-l mailing list
Wikisource-l@lists.wikimedia.org
Wikisource-l mailing list
Wikisource-l@lists.wikimedia.org
-----Inline Attachment Follows-----
Wikisource-l mailing list
Wikisource-l@lists.wikimedia.org
Wikisource-l mailing list
Wikisource-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikisource-l
Wikisource-l mailing list
Wikisource-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikisource-l
-----Inline Attachment Follows-----
Wikisource-l mailing list Wikisource-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikisource-l