Re: [Wikitech-l] [Foundation-l] Wikitext isn't aprogramminglanguage, but...

classic Classic list List threaded Threaded
17 messages Options
Reply | Threaded
Open this post in threaded view
|

Re: [Wikitech-l] [Foundation-l] Wikitext isn't aprogramminglanguage, but...

Bugzilla from andrew@epstone.net

Ligulem wrote:
> On 19.01.2007 10:27, Andrew Garrett wrote:
> > If you want something implemented in a pure programming language
> > (PHP is debatable in that regard), speak to developers, we will
> > implement it in PHP as part of MediaWiki. MediaWiki is a piece
> > of software, not an IDE.
>
> Would you move http://en.wikipedia.org/wiki/Template:Cite_book into
> MediaWiki then?
>
Taking this seriously, instead of as bait, if you are serious about
this, please post a bugzilla request, instead of hounding individual
developers.

That said, I think that a basic parameterised thing like that
needn't be reimplemented; although it would be perhaps helpful in
performance terms.

Andrew Garrett
(Werdna)

_______________________________________________
Wikitech-l mailing list
[hidden email]
http://lists.wikimedia.org/mailman/listinfo/wikitech-l
Reply | Threaded
Open this post in threaded view
|

Re: [Wikitech-l] [Foundation-l] Wikitext isn't aprogramminglanguage, but...

Ligulem
On 19.01.2007 11:55, Andrew Garrett wrote:
> Taking this seriously, instead of as bait, if you are serious about
> this, please post a bugzilla request, instead of hounding individual
> developers.
>

Hounding?

Thanks a lot for assuming bad faith.






_______________________________________________
Wikitech-l mailing list
[hidden email]
http://lists.wikimedia.org/mailman/listinfo/wikitech-l
Reply | Threaded
Open this post in threaded view
|

Re: [Wikitech-l] [Foundation-l] Wikitext isn't aprogramminglanguage, but...

Anthony-61
In reply to this post by Bugzilla from andrew@epstone.net
On 1/19/07, Andrew Garrett <[hidden email]> wrote:

>
> Ligulem wrote:
> > On 19.01.2007 10:27, Andrew Garrett wrote:
> > > If you want something implemented in a pure programming language
> > > (PHP is debatable in that regard), speak to developers, we will
> > > implement it in PHP as part of MediaWiki. MediaWiki is a piece
> > > of software, not an IDE.
> >
> > Would you move http://en.wikipedia.org/wiki/Template:Cite_book into
> > MediaWiki then?
> >
> Taking this seriously, instead of as bait, if you are serious about
> this, please post a bugzilla request, instead of hounding individual
> developers.
>
You said "speak to developers", not fill out a bugzilla request.  Now,
if one of us *does* put in a bugzilla request, will it get done?

Unfortunately, the MediaWiki PHP code isn't as easy to update as
templates are.  Of course, resolving this without causing major
security headaches is/would be an insanely difficult thing to do.

> That said, I think that a basic parameterised thing like that
> needn't be reimplemented; although it would be perhaps helpful in
> performance terms.
>
If it were implemented in PHP, could then a lookup be done so one
would only need to give the ISBN and page number, and the rest would
be added automagically?

_______________________________________________
Wikitech-l mailing list
[hidden email]
http://lists.wikimedia.org/mailman/listinfo/wikitech-l
Reply | Threaded
Open this post in threaded view
|

Re: [Wikitech-l] [Foundation-l] Wikitext isn't aprogramminglanguage, but...

Rob Church
On 19/01/07, Anthony <[hidden email]> wrote:
> If it were implemented in PHP, could then a lookup be done so one
> would only need to give the ISBN and page number, and the rest would
> be added automagically?

Where would we get the data from? Who's going to allow us to query
this stuff at the rate we need it? Sure, we can cache it, but it still
takes time to build up a good cache of titles.

Coincidentally, I just finished the beta version of an extension,
BookInformation, which enhances the output of Special:Booksources with
a cute little box containing the title, author and publisher
corresponding to the ISBN entered, where it exists. Works with Amazon
Web Services right now, but has been coded to allow a number of
possible providers to be supported. (Let me take the opportunity to
thank the guys who shot in and provided translations for Finnish,
Dutch, German and Indonesian within 10 minutes of committing it; if
that's not bloody service, I don't know what is!)

There's no reason we couldn't enhance the citation extension with an
auto-formatter for book/journal/similar references, *but* consider
that to do this, we'd be pulling the format from a message as it is;
so there isn't much benefit over using a template as is done at
present.


Rob Church

_______________________________________________
Wikitech-l mailing list
[hidden email]
http://lists.wikimedia.org/mailman/listinfo/wikitech-l
Reply | Threaded
Open this post in threaded view
|

Re: [Wikitech-l] [Foundation-l] Wikitext isn't aprogramminglanguage, but...

Simetrical
On 1/19/07, Rob Church <[hidden email]> wrote:

> On 19/01/07, Anthony <[hidden email]> wrote:
> > If it were implemented in PHP, could then a lookup be done so one
> > would only need to give the ISBN and page number, and the rest would
> > be added automagically?
>
> . . .
>
> There's no reason we couldn't enhance the citation extension with an
> auto-formatter for book/journal/similar references, *but* consider
> that to do this, we'd be pulling the format from a message as it is;
> so there isn't much benefit over using a template as is done at
> present.

Current templates require you to specify all the information to be
used in the templates.  The question related to supplying only the
ISBN and having it come up with the rest of the information from its
database, as well as formatting it properly.

_______________________________________________
Wikitech-l mailing list
[hidden email]
http://lists.wikimedia.org/mailman/listinfo/wikitech-l
Reply | Threaded
Open this post in threaded view
|

Re: [Wikitech-l] [Foundation-l] Wikitext isn't aprogramminglanguage, but...

Rob Church
On 19/01/07, Simetrical <[hidden email]> wrote:
> Current templates require you to specify all the information to be
> used in the templates.  The question related to supplying only the
> ISBN and having it come up with the rest of the information from its
> database, as well as formatting it properly.

The kickstart for the post was actually reimplementing
Template:Cite_book or whatever; the ISBN lookup stuff was added later.


Rob Church

_______________________________________________
Wikitech-l mailing list
[hidden email]
http://lists.wikimedia.org/mailman/listinfo/wikitech-l
Reply | Threaded
Open this post in threaded view
|

Re: [Wikitech-l] [Foundation-l] Wikitext isn't aprogramminglanguage, but...

Simetrical
On 1/19/07, Rob Church <[hidden email]> wrote:
> The kickstart for the post was actually reimplementing
> Template:Cite_book or whatever; the ISBN lookup stuff was added later.

But that isn't the post you were responding to, and anyway the poster
brings up a good point.

_______________________________________________
Wikitech-l mailing list
[hidden email]
http://lists.wikimedia.org/mailman/listinfo/wikitech-l
Reply | Threaded
Open this post in threaded view
|

Re: [Wikitech-l] [Foundation-l] Wikitext isn't aprogramminglanguage, but...

Rob Church
On 19/01/07, Simetrical <[hidden email]> wrote:
> But that isn't the post you were responding to, and anyway the poster
> brings up a good point.

I didn't say he hadn't!

Do *you* have any solutions to the main problem I pointed out?


Rob Church

_______________________________________________
Wikitech-l mailing list
[hidden email]
http://lists.wikimedia.org/mailman/listinfo/wikitech-l
Reply | Threaded
Open this post in threaded view
|

Re: [Wikitech-l] [Foundation-l] Wikitext isn't aprogramminglanguage, but...

Anthony-61
In reply to this post by Rob Church
On 1/19/07, Rob Church <[hidden email]> wrote:
> On 19/01/07, Anthony <[hidden email]> wrote:
> > If it were implemented in PHP, could then a lookup be done so one
> > would only need to give the ISBN and page number, and the rest would
> > be added automagically?
>
> Where would we get the data from? Who's going to allow us to query
> this stuff at the rate we need it? Sure, we can cache it, but it still
> takes time to build up a good cache of titles.
>
How fast of a rate would you need it?  I don't know the answer to
where the data could be obtained, that's why I asked the initial
question about whether or not this would be possible.

Are these databases copyrightable?  Can the Library of Congress
database be obtained with a Freedom of Information Request?

Anthony

_______________________________________________
Wikitech-l mailing list
[hidden email]
http://lists.wikimedia.org/mailman/listinfo/wikitech-l
Reply | Threaded
Open this post in threaded view
|

Re: [Wikitech-l] [Foundation-l] Wikitext isn't aprogramminglanguage, but...

Rob Church
On 19/01/07, Anthony <[hidden email]> wrote:
> How fast of a rate would you need it?  I don't know the answer to
> where the data could be obtained, that's why I asked the initial
> question about whether or not this would be possible.

It'd need to be available at rendering time; if a cache miss occurred,
it would need to be queried there and then. I don't know how fast
people would adopt it or how many such tags would be inserted in a
short period of time, so estimating the worst case is a bit hard.

I'm not sure how "available" such databases would be.


Rob Church

_______________________________________________
Wikitech-l mailing list
[hidden email]
http://lists.wikimedia.org/mailman/listinfo/wikitech-l
Reply | Threaded
Open this post in threaded view
|

Re: [Wikitech-l] [Foundation-l] Wikitext isn't aprogramminglanguage, but...

Edward Z. Yang
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Rob Church wrote:
> It'd need to be available at rendering time; if a cache miss occurred,
> it would need to be queried there and then. I don't know how fast
> people would adopt it or how many such tags would be inserted in a
> short period of time, so estimating the worst case is a bit hard.

Don't forget that you could also stuff it into the job queue. There
would be a loss immediate feedback, but eventually the correct data
would be cached and then re-running the parser would make things work out.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (MingW32)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFFsURpqTO+fYacSNoRAlrzAJ4ixdgyeUwa8K+EJCvAYUCFtUxkEQCfU3yl
ia7y/wVgq8+gYFoPH/VsxPI=
=ZIVj
-----END PGP SIGNATURE-----


_______________________________________________
Wikitech-l mailing list
[hidden email]
http://lists.wikimedia.org/mailman/listinfo/wikitech-l
Reply | Threaded
Open this post in threaded view
|

Re: [Wikitech-l] [Foundation-l] Wikitext isn't aprogramminglanguage, but...

Anthony-61
In reply to this post by Rob Church
On 1/19/07, Rob Church <[hidden email]> wrote:
> On 19/01/07, Anthony <[hidden email]> wrote:
> > How fast of a rate would you need it?  I don't know the answer to
> > where the data could be obtained, that's why I asked the initial
> > question about whether or not this would be possible.
>
> It'd need to be available at rendering time; if a cache miss occurred,
> it would need to be queried there and then.

Umm, not if you designed it so the "cache" was big enough.  Then you'd
only need to perform the lookup once - presumably when the reference
is first added.

> I don't know how fast
> people would adopt it or how many such tags would be inserted in a
> short period of time, so estimating the worst case is a bit hard.
>
I can't possibly answer your question of "Who's going to allow us to
query this stuff at the rate we need it?" then.

_______________________________________________
Wikitech-l mailing list
[hidden email]
http://lists.wikimedia.org/mailman/listinfo/wikitech-l
Reply | Threaded
Open this post in threaded view
|

Re: [Wikitech-l] [Foundation-l] Wikitext isn't aprogramminglanguage, but...

Andrew Dunbar
In reply to this post by Rob Church
On 1/20/07, Rob Church <[hidden email]> wrote:
> On 19/01/07, Anthony <[hidden email]> wrote:
> > If it were implemented in PHP, could then a lookup be done so one
> > would only need to give the ISBN and page number, and the rest would
> > be added automagically?
>
> Where would we get the data from? Who's going to allow us to query
> this stuff at the rate we need it? Sure, we can cache it, but it still
> takes time to build up a good cache of titles.

Try librarything.com - they have an API and were keen recently for people
to do mashups. Tell Tim I recommended it (-;

Andrew Dunbar (hippietrail)

> Coincidentally, I just finished the beta version of an extension,
> BookInformation, which enhances the output of Special:Booksources with
> a cute little box containing the title, author and publisher
> corresponding to the ISBN entered, where it exists. Works with Amazon
> Web Services right now, but has been coded to allow a number of
> possible providers to be supported. (Let me take the opportunity to
> thank the guys who shot in and provided translations for Finnish,
> Dutch, German and Indonesian within 10 minutes of committing it; if
> that's not bloody service, I don't know what is!)
>
> There's no reason we couldn't enhance the citation extension with an
> auto-formatter for book/journal/similar references, *but* consider
> that to do this, we'd be pulling the format from a message as it is;
> so there isn't much benefit over using a template as is done at
> present.
>
>
> Rob Church
>
> _______________________________________________
> Wikitech-l mailing list
> [hidden email]
> http://lists.wikimedia.org/mailman/listinfo/wikitech-l
>


--
http://linguaphile.sf.net

_______________________________________________
Wikitech-l mailing list
[hidden email]
http://lists.wikimedia.org/mailman/listinfo/wikitech-l
Reply | Threaded
Open this post in threaded view
|

Re: [Wikitech-l] [Foundation-l] Wikitext isn't aprogramminglanguage, but...

Rob Church
In reply to this post by Anthony-61
On 19/01/07, Anthony <[hidden email]> wrote:
> Umm, not if you designed it so the "cache" was big enough.  Then you'd
> only need to perform the lookup once - presumably when the reference
> is first added.

I should clarify that by "cache miss", I meant, "the first use, when
it won't be in the cache" - of course, we wouldn't arbitrarily discard
the cached data; there's no need.


Rob Church

_______________________________________________
Wikitech-l mailing list
[hidden email]
http://lists.wikimedia.org/mailman/listinfo/wikitech-l
Reply | Threaded
Open this post in threaded view
|

Re: [Wikitech-l] [Foundation-l] Wikitext isn't aprogramminglanguage, but...

Jens Frank
In reply to this post by Rob Church
On Fri, Jan 19, 2007 at 08:00:17PM +0000, Rob Church wrote:
> On 19/01/07, Anthony <[hidden email]> wrote:
> > If it were implemented in PHP, could then a lookup be done so one
> > would only need to give the ISBN and page number, and the rest would
> > be added automagically?
>
> Where would we get the data from? Who's going to allow us to query
> this stuff at the rate we need it? Sure, we can cache it, but it still
> takes time to build up a good cache of titles.

After some googling, I found

        http://isbndb.com/docs/api/

Would this be helpful? They claim to have about 2.5 Million titles. The
search found some O'Reilly books that I tried, but failed on some German
titles I tried.

Regards,

        jens

_______________________________________________
Wikitech-l mailing list
[hidden email]
http://lists.wikimedia.org/mailman/listinfo/wikitech-l
Reply | Threaded
Open this post in threaded view
|

Re: [Wikitech-l] [Foundation-l] Wikitext isn't aprogramminglanguage, but...

Rob Church
On 20/01/07, Jens Frank <[hidden email]> wrote:
> After some googling, I found
>
>         http://isbndb.com/docs/api/
>
> Would this be helpful? They claim to have about 2.5 Million titles. The
> search found some O'Reilly books that I tried, but failed on some German
> titles I tried.

Good find. It might be advisable for us to compile a list of such
providers - if we can establish common data to be obtained for each
title, then perhaps we can do some sort of basic round-robin querying,
rather than pulling all our information from one provider.

Thanks,


Rob Church

_______________________________________________
Wikitech-l mailing list
[hidden email]
http://lists.wikimedia.org/mailman/listinfo/wikitech-l
Reply | Threaded
Open this post in threaded view
|

Re: [Wikitech-l] [Foundation-l] Wikitext isn't aprogramminglanguage, but...

Jay Ashworth-2
In reply to this post by Anthony-61
On Fri, Jan 19, 2007 at 04:52:56PM -0500, Anthony wrote:
> Are these databases copyrightable?  Can the Library of Congress
> database be obtained with a Freedom of Information Request?

Databases are inherently not copyrightable (Feist v. Rural).

The LoC is operated for our benefit anyway; I would be very surprised
if they didn't already have some fashion by which you could look stuff
up over the net.

This may be a good start: http://www.loc.gov/library/

(for anyone who hasn't already seen it...)

Cheers,
-- jra
--
Jay R. Ashworth                                                [hidden email]
Designer                          Baylink                             RFC 2100
Ashworth & Associates        The Things I Think                        '87 e24
St Petersburg FL USA      http://baylink.pitas.com             +1 727 647 1274

_______________________________________________
Wikitech-l mailing list
[hidden email]
http://lists.wikimedia.org/mailman/listinfo/wikitech-l