Shorter Url for non-latin languages

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

Shorter Url for non-latin languages

mingli yuan
Hi, folks,

About four months ago, I post a message about a url shortening service for
Wikipedias, esp. the non-latin language version.
Yesterday, I moved the service to a new domain http://defn.me , and also
updated related user-scripts in en, ml, ru, ta, he, zh.

* http://en.wikipedia.org/wiki/User:Mountain/shorturl
* http://en.wikipedia.org/wiki/User:Mountain/shorturl.js

This is a very small service, but I am quite sure this service is useful
for non-latin languages.
This work is some kind of usability improvement.

So I just want to know the possibility that foundation can support it or
not?
And how should I improve the work to make foundation accept this service?

Thanks.

Regards,
Mingli (User:Mountain)
_______________________________________________
foundation-l mailing list
[hidden email]
Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/foundation-l
Reply | Threaded
Open this post in threaded view
|

Re: Shorter Url for non-latin languages

David Gerard-2
On 11 February 2011 11:30, Mingli Yuan <[hidden email]> wrote:

> So I just want to know the possibility that foundation can support it or
> not?
> And how should I improve the work to make foundation accept this service?


A URL shortener is a very good idea.

Even for English, there's http://enwp.org for instance. English
Wikinews has http://enwn.net . Neither of these is official.

But maybe having an official shortener would be a good idea.


- d.

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

Re: Shorter Url for non-latin languages

Gerard Meijssen-3
Hoi,
Even better then URL shorteners is the presentation of the URL in Unicode.
This will make the URL readable, memorable and guessable. It may be that
this has everything to do with the browser that is used.. It is certainly a
good reason to select one browser over an other.

http://ru.wikipedia.org/wiki/%D0%97%D0%B0%D0%B3%D0%BB%D0%B0%D0%B2%D0%BD%D0%B0%D1%8F_%D1%81%D1%82%D1%80%D0%B0%D0%BD%D0%B8%D1%86%D0%B0shows
properly in my browser, it gets spoiled when I past it in my mail..
Thanks,
      GerardM

On 11 February 2011 14:36, David Gerard <[hidden email]> wrote:

> On 11 February 2011 11:30, Mingli Yuan <[hidden email]> wrote:
>
> > So I just want to know the possibility that foundation can support it or
> > not?
> > And how should I improve the work to make foundation accept this service?
>
>
> A URL shortener is a very good idea.
>
> Even for English, there's http://enwp.org for instance. English
> Wikinews has http://enwn.net . Neither of these is official.
>
> But maybe having an official shortener would be a good idea.
>
>
> - d.
>
> _______________________________________________
> foundation-l mailing list
> [hidden email]
> Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/foundation-l
>
_______________________________________________
foundation-l mailing list
[hidden email]
Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/foundation-l
Reply | Threaded
Open this post in threaded view
|

Re: Shorter Url for non-latin languages

MZMcBride-2
In reply to this post by David Gerard-2
David Gerard wrote:
> On 11 February 2011 11:30, Mingli Yuan <[hidden email]> wrote:
>> So I just want to know the possibility that foundation can support it or not?
>> And how should I improve the work to make foundation accept this service?
>
> A URL shortener is a very good idea.

In general, URL shorteners are a terrible idea. They often rely on
third-party services, so at any point the links could put a pay wall or ads
in between click and target (or worse, stop working completely). They also
generally greatly reduce the value of a URL. They're good for spammers,
though.

> Even for English, there's http://enwp.org for instance. English
> Wikinews has http://enwn.net . Neither of these is official.

These are a little better because even if enwp.org dies or becomes a
different kind of site, it's possible to know what someone intended when
they wrote "enwp.org/foo".

> But maybe having an official shortener would be a good idea.

Maybe. It would at least mitigate the risk of a third-party going belly up.
A lot of organizations are using their own short URLs for this reason
(nytim.es and the like). That said, every page has a page ID (curid) and a
revision ID (oldid):

* http://en.wikipedia.org/wiki/Metasyntactic_variable
* http://en.wikipedia.org/w/index.php?oldid=411553662
* http://en.wikipedia.org/w/index.php?curid=20036

These can be made even shorter:
* http://en.wikipedia.org/?oldid=411553662
* http://en.wikipedia.org/?curid=20036

Drawbacks to this are that using IDs instead of titles decreases the value
of the URLs and "curid" (a reference to the internal page.page_id) can
change if a page is deleted/undeleted. There are also interactions with
redirects and page moves to consider.

It largely depends on what the use-case for having such a short URL is going
to be and how many costs are worth those benefits. The use-cases still seem
rather confined to me, while the overhead to setting up and maintaining such
a service is not negligible.

MZMcBride



_______________________________________________
foundation-l mailing list
[hidden email]
Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/foundation-l
FT2
Reply | Threaded
Open this post in threaded view
|

Re: Shorter Url for non-latin languages

FT2
Shorteners are one way that a large part of the net is going.

Twitter and other microblogging drives that - a full WMF url won't easily
fit so people probably don't include them in many media where they otherwise
might well do so.

Making it clear these are WMF related, and also hosting them and/or having
rights to manage them,  is important. Handling permalinks and not just
current revisions is important.

wmf.org/* may or may not be a good one, but if not, then other possibilities
that convey the identity of the project can surely be identified if needed.
Not going to suggest any others on the open mailing list, but if the idea is
worth trying, I'm sure we'll find out to whom to send suggestions.

FT2


On Fri, Feb 11, 2011 at 11:56 PM, MZMcBride <[hidden email]> wrote:

> David Gerard wrote:
> > On 11 February 2011 11:30, Mingli Yuan <[hidden email]> wrote:
> >> So I just want to know the possibility that foundation can support it or
> not?
> >> And how should I improve the work to make foundation accept this
> service?
> >
> > A URL shortener is a very good idea.
>
> In general, URL shorteners are a terrible idea. They often rely on
> third-party services, so at any point the links could put a pay wall or ads
> in between click and target (or worse, stop working completely). They also
> generally greatly reduce the value of a URL. They're good for spammers,
> though.
>
> > Even for English, there's http://enwp.org for instance. English
> > Wikinews has http://enwn.net . Neither of these is official.
>
> These are a little better because even if enwp.org dies or becomes a
> different kind of site, it's possible to know what someone intended when
> they wrote "enwp.org/foo".
>
> > But maybe having an official shortener would be a good idea.
>
> Maybe. It would at least mitigate the risk of a third-party going belly up.
> A lot of organizations are using their own short URLs for this reason
> (nytim.es and the like). That said, every page has a page ID (curid) and a
> revision ID (oldid):
>
> * http://en.wikipedia.org/wiki/Metasyntactic_variable
> * http://en.wikipedia.org/w/index.php?oldid=411553662
> * http://en.wikipedia.org/w/index.php?curid=20036
>
> These can be made even shorter:
> * http://en.wikipedia.org/?oldid=411553662
> * http://en.wikipedia.org/?curid=20036
>
> Drawbacks to this are that using IDs instead of titles decreases the value
> of the URLs and "curid" (a reference to the internal page.page_id) can
> change if a page is deleted/undeleted. There are also interactions with
> redirects and page moves to consider.
>
> It largely depends on what the use-case for having such a short URL is
> going
> to be and how many costs are worth those benefits. The use-cases still seem
> rather confined to me, while the overhead to setting up and maintaining
> such
> a service is not negligible.
>
> MZMcBride
>
>
>
> _______________________________________________
> foundation-l mailing list
> [hidden email]
> Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/foundation-l
>
_______________________________________________
foundation-l mailing list
[hidden email]
Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/foundation-l
Reply | Threaded
Open this post in threaded view
|

Re: Shorter Url for non-latin languages

Alison M. Wheeler
In reply to this post by mingli yuan
I'm not following exactly who said
> > A URL shortener is a very good idea.
but I must firmly agree.

On Fri, Feb 11, 2011 at 11:56 PM, MZMcBride <[hidden email]> wrote:
> It largely depends on what the use-case for having such a short URL is
> going to be and how many costs are worth those benefits. The use-cases
> still seem rather confined to me, while the overhead to setting up and
> maintaining such a service is not negligible.

Twitter and its ilk (identi.ca etc) have set an expectation that people can communicate in 140 characters* and if a user currently wants to point to an article / version of ours then they will already be using a shortening service outwith our control. The use case is already well-defined and in use.

We can't stop them doing so, but we could assist them and, in so doing, improve our services.

As David and others have pointed out, if other services continue to provide this service then they get to capture information about what is 'hot', and if they go under the links become worthless. This applies even if the world has moved on (ime most short links have a lifetime measured in weeks or even days).

As regards setting up and maintaining such a service I am happy to say that the cost *is* entirely negligible. Because I was loathe to let some search engine see what targets I was storing (and because other people's sites do disappear) I wrote my own some time back. Didn't take too long and - unlike the bit.ly's of this world - when a link target is submitted it first checks if it already exists, thus removing the duplication which is so common amongst current services.

I'd be happy to make the code open (CC-BY or GPL3 probably), but it isn't that difficult, or expensive in processing / storage terms.

An alternative would be to create the link (in effect an unlabelled redirect) from a short domain (http://wp.wmf would be nice!) for every page and revision automatically, and include it on the page, thus saving on any database requirement because of the 1:1 matching.

The best option, naturally, would be to get Twitter and everyone else to accept "[[target]]" as an automated link! (but that would still be really too long for most pages)

Alison


* UTF codepoints, to be precise


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

Re: Shorter Url for non-latin languages

Alison M. Wheeler

I forgot to add, in my belief we should be offering a URL shortening service / route for *all* languages, not just the non-Latin charset ones.

AlisonW

(ps. my service has been running for over 2 1/2 years and is completely set-and-forget; I never need to look at it!)

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

Re: Shorter Url for non-latin languages

Bugzilla from hofmanj@aldebaran.cz
In reply to this post by Alison M. Wheeler
> > > A URL shortener is a very good idea.

I like the idea.

Very nice url would be for example: http://en.wp.wmf.org/Az09Q . This would make possible to see which project the link leads to.

Since (26*2+10)**5 = 916132832 , five chars should be enough for all articles for ever.

If some universal url shortener was created, for example http://wmf.org/ , then 7 chars (3521614606208) would be enough for all projects. If not, the 8th char can always be added.

Jiri

_______________________________________________
foundation-l mailing list
[hidden email]
Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/foundation-l

signature.asc (205 bytes) Download Attachment
FT2
Reply | Threaded
Open this post in threaded view
|

Re: Shorter Url for non-latin languages

FT2
People will want to link to permalinks.

8 chars, 64 character set (a-z A-Z 0-9 and 2 others). Unlikely to run out
for the foreseeable future.

FT2


On Sat, Feb 12, 2011 at 11:26 AM, Jiri Hofman <[hidden email]> wrote:

> Since (26*2+10)**5 = 916132832 , five chars should be enough for all
> articles for ever.
>
> If some universal url shortener was created, for example http://wmf.org/ ,
> then 7 chars (3521614606208) would be enough for all projects. If not, the
> 8th char can always be added.
>
> Jiri
>
_______________________________________________
foundation-l mailing list
[hidden email]
Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/foundation-l
Reply | Threaded
Open this post in threaded view
|

Re: Shorter Url for non-latin languages

The Mono
In reply to this post by Bugzilla from hofmanj@aldebaran.cz
Of course, this is not possible.

> Very nice url would be for example: http://en.wp.wmf.org/Az09Q . This would make possible to see which project the link leads to.

Wmf.org is already registered, but in the future, a .Wmf TLD might be possible.


--
*Mono*
http://enwp.org/m:User:Mono

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

Re: Shorter Url for non-latin languages

Ryu Cheol
There is a related proposal at strategy :
http://strategy.wikimedia.org/wiki/Proposal_talk:.WIKI._and_.WK._top_level_domains

<http://strategy.wikimedia.org/wiki/Proposal_talk:.WIKI._and_.WK._top_level_domains>I
think we can choose our own pathname  manually just like
http://en.wp.wmf.org/WMF. (Wikimedia Foundation)
And When you want to point out exact paragraph,
http://en.wp.wmf.org/WMF#1.2(We usually do not use numbers as
paragraph headers, in this case second
section of a first paragraph)

On the proposal, I also suggested to add twitter button on every articles.
 Those shortened URLs might be helpful.

Cheol

2011/2/13 The Mono <[hidden email]>

> Of course, this is not possible.
>
> > Very nice url would be for example: http://en.wp.wmf.org/Az09Q . This
> would make possible to see which project the link leads to.
>
> Wmf.org is already registered, but in the future, a .Wmf TLD might be
> possible.
>
>
> --
> *Mono*
> http://enwp.org/m:User:Mono
>
> _______________________________________________
> foundation-l mailing list
> [hidden email]
> Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/foundation-l
>
_______________________________________________
foundation-l mailing list
[hidden email]
Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/foundation-l
Reply | Threaded
Open this post in threaded view
|

Re: Shorter Url for non-latin languages

theo10011
I thought the biggest reason to get a url shortener was suggested as links
in and from non-latin languages, the issue was character encoding for
non-latin scripts.

But if we're considering top level domains already, how about our own tld
for all the projects. The foundation already has hundreds of projects, a
single tld for all current and future project- .wmf or .wiki.

Theo

On Sun, Feb 13, 2011 at 12:55 AM, RYU Cheol <[hidden email]> wrote:

> There is a related proposal at strategy :
>
> http://strategy.wikimedia.org/wiki/Proposal_talk:.WIKI._and_.WK._top_level_domains
>
> <
> http://strategy.wikimedia.org/wiki/Proposal_talk:.WIKI._and_.WK._top_level_domains
> >I
> think we can choose our own pathname  manually just like
> http://en.wp.wmf.org/WMF. (Wikimedia Foundation)
> And When you want to point out exact paragraph,
> http://en.wp.wmf.org/WMF#1.2(We usually do not use numbers as
> paragraph headers, in this case second
> section of a first paragraph)
>
> On the proposal, I also suggested to add twitter button on every articles.
>  Those shortened URLs might be helpful.
>
> Cheol
>
> 2011/2/13 The Mono <[hidden email]>
>
> > Of course, this is not possible.
> >
> > > Very nice url would be for example: http://en.wp.wmf.org/Az09Q . This
> > would make possible to see which project the link leads to.
> >
> > Wmf.org is already registered, but in the future, a .Wmf TLD might be
> > possible.
> >
> >
> > --
> > *Mono*
> > http://enwp.org/m:User:Mono
> >
> > _______________________________________________
> > foundation-l mailing list
> > [hidden email]
> > Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/foundation-l
> >
> _______________________________________________
> foundation-l mailing list
> [hidden email]
> Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/foundation-l
>
_______________________________________________
foundation-l mailing list
[hidden email]
Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/foundation-l
FT2
Reply | Threaded
Open this post in threaded view
|

Re: Shorter Url for non-latin languages

FT2
In reply to this post by Ryu Cheol
Yes x 2.

But keep the links simple, so extended subdomains (en.wp.wmf.org) is too
long. Just one domain, and let the shortener cover all wikis not just one.

FT2


On Sat, Feb 12, 2011 at 11:08 AM, Alison M. Wheeler <
[hidden email]> wrote:

> An alternative would be to create the link (in effect an unlabelled
> redirect) from a short domain (http://wp.wmf would be nice!) for every
> page and revision automatically, and include it on the page, thus saving on
> any database requirement because of the 1:1 matching.
> The best option, naturally, would be to get Twitter and everyone else to
> accept "[[target]]" as an automated link! (but that would still be really
> too long for most pages)
>



On Sat, Feb 12, 2011 at 7:25 PM, RYU Cheol <[hidden email]> wrote:

> On the proposal, I also suggested to add twitter button on every articles.
>  Those shortened URLs might be helpful.
>
_______________________________________________
foundation-l mailing list
[hidden email]
Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/foundation-l