marnanel: (party hat)Monument ([personal profile] marnanel) wrote,
@ 2010-12-17 01:35 pm UTC
  • Previous Entry
  • Add to Memories
  • Tell someone about this!
  • Next Entry
Entry tags:code, pccu, pgo
There has been a lot of fuss about del.icio.us shutting down (along with AltaVista— my goodness, I used that when I was an undergrad. How weird to see it go).

I am worried that someone out there is going to go to all the effort of writing a new del.ici.ous. Or more likely several someones, and all the results will be incompatible. As Mark Pilgrim points out, this is a bad idea. Don't do it.

I'd like to offer a better suggestion. Write something that takes the not-really-XML of the del.icio.us backup format and uploads it line by line to a given identi.ca account. So, for example,

<DT><A HREF="http://www.paulmeier.com/ipa/charts.html" ADD_DATE="1104124425" PRIVATE="0" TAGS="flash,linguistics,phonetics">Flash-based IPA charts</A>

becomes

http://www.paulmeier.com/ipa/charts.html Flash-based IPA charts #flash #linguistics #phonetics

That way
  • you keep your bookmarks somewhere
  • that somewhere runs on free software
  • tags still work
  • you can make feeds
  • you can get daily dumps added to your blog
and so on.  If there's something it can't do, it's far better to do it in identi.ca and then extend identi.ca than to write a whole new site.

The main issue is that the string becomes longer than 140 characters, but there are ways around that.

If anyone writes such a script, let me know and I'll link it here.

(7 comments) - (Post a new comment)
(Flat) (Top-level comments only)

Why not Weave?


[identity profile] adamwill.id.fedoraproject.org
2010-12-17 10:10 pm UTC (link)
Why not something that just dumps them into Weave / Firefox Sync?

I'm going to keep bashing on about Firefox Sync until everyone twigs that it's awesome: it's the cloud done right. (Even though it's called Firefox Sync there's no reason it needs to be tied to Firefox / Mozilla - it's a completely open and documented protocol and the official client and server bits are of course both open source and hence should be useful to write code for other things to hook into the system).

(Reply to this)  (Thread) 


marnanel: (party hat)

Re: Why not Weave?


[personal profile] marnanel
2010-12-18 02:09 am UTC (link)
I didn't mention Weave and Firefox Sync mostly because I know nothing about them. :)

Do you have a good introduction handy?

(Reply to this)  (Thread from start)  (Parent) 


Space


(Anonymous)
2010-12-18 01:48 am UTC (link)
"The main issue is that the string becomes longer than 140 characters, but there are ways around that."

How exactly?

A tool for a problem, nothing more. Identica/Twitter barely serve the purpose that they where built for.

Someone else suggested Facebook. Ugh! lets all commit suicide.

(Reply to this)  (Thread) 


marnanel: (party hat)

Re: Space


[personal profile] marnanel
2010-12-18 02:08 am UTC (link)
First pass: using a URL shortener.

More complex solution: using the identi.ca codebase, but removing the limit of 140 characters.

Your unreasoned dismissal of the entire idea behind Twitter and identi.ca does not give me much confidence in your other thoughts.

(Reply to this)  (Thread from start)  (Parent) 


Psudo.


[identity profile] rustycurry.pip.verisignlabs.com
2010-12-18 06:47 am UTC (link)
The reason to put it on identi.ca is rather than just a sync tool is to make sure that others can find the information later. That was the reason to put it on del.icio.us in the first place. But if sync works for you go with it.

I would think that while it's not quite xml spec, it is one of the html specs. as a result the python html library should be able to parse the content.

Presuming you can walk the content of the backup page for links, (
[Error: Irreparable invalid markup ('<a [...] ../a>') in entry. Owner must fix manually. Raw contents below.]

The reason to put it on identi.ca is rather than just a sync tool is to make sure that others can find the information later. That was the reason to put it on del.icio.us in the first place. But if sync works for you go with it.

I would think that while it's not quite xml spec, it is one of the html specs. as a result the python html library should be able to parse the content.

Presuming you can walk the content of the backup page for links, (<A HREF ../A> blocks) the parser should be able to give you the HREF, TAGS and description portions of the link as variables to handle. You then need to basically walk through the tags building a tag_string with each tag word preceded by a ' #' string.

Then you build a post_string as url + description + tag_string. Check to see if the length of the post_string is less than or equal to 140 character, if it is post using the identi.ca libraries for python, or through any tool you have automated already. (I would use a python script that would post through my xmlpp server, but that's a different topic.)

If the post_string is greater than 140 find out if you can run it through a url shortener and get something shorter. If it's already a shortened url, you may want to have it spit out the string to a separate file to go through on your own to see what you can do about shortening the description, or perhaps discarding or shortening tags.

(Reply to this) 


The best out there


[identity profile] sspaeth.de
2010-12-18 02:26 pm UTC (link)
While I usually don't plug commercial non-open source sites: pinboard.in is really phantastic. It has a delicious compatible API, so if you used applications to put/access your URLs, it will continue to work. It can automatically suck in you delicious URLs and provides nice bookmarking bookmarklets. Lifetime sign up cost is <7$. I like it a lot.

(Reply to this) 


Here you go


(Anonymous)
2011-01-06 03:17 am UTC (link)
https://github.com/thekad/deliciousness

(Reply to this) 


(7 comments) - (Post a new comment)
(Flat) (Top-level comments only)