API use cases

[maciej]

For those of you who have asked about API access, can you post some
use cases to help in the design?

Despite my fear of SSL configuration I am leaning towards requiring
https://, so I'm particularly interested to hear if there are any use
cases that can't deal with that protocol.

[TylerWeir]

I would love to build a desktop and iPhone app for Pinboard.

With the addition of an API I would hope that the authors of twitter
apps would include "post to pinboard" when URLs are detected. I do
that manually right now.

I'm sure someone has a more creative idea, as mine are a bit
utilitarian.

Tyler

[Ramanan]

> For those of you who have asked about API access, can you post some
> use cases to help in the design?

Posting a link was the big thing I was looking to do. I used to have a
script I would use to link to stuff on my blog, and cross post it to
pinboard. This would also let people writing Twitter iPhone clients
and other junk like that to post to a users pinboard account.

One thing I want to do now is write a script that will pull posts
tagged with something like "posttoblog" from pinboard and post them to
my blog, but I can probably do this with the RSS feed.

[Romain]

I've used a combination of Quicksilver and delicious for 2 or 3 years
now.

It's so comfortable… it actually kind of prevented me to use pinboard
full speed.
Now i have a very dirty PHP script that pulls the exported XML and
transforms it into an HTML that can be parsed by QS.

That will work for some time.

Having an API would greatly help whoever is more skilled than me to
create a QS plug-in.

[seanh]

On Mon, Jul 27, 2009 at 3:00 AM, maciej<mcegl...@gmail.com> wrote:
>
> For those of you who have asked about API access, can you post some
> use cases to help in the design?

I'd like to write a command-line client, writing in Python, that would
be able to connect to a pinboard account and do things like:

* Download the most recent bookmarks
* Download your list of tags
* Download all bookmarks with a given tag
* Download all bookmarks matching a given search query
* Download starred items
* Download to read items
* Edit bookmark names and descriptions, tags, star and unstar
bookmarks, mark them read

So basically everything you can do through the web interface, except
that actually using a command-line client to add a new bookmark seems
silly.

You could use this to print out your bookmarks to the command line,
but you could also save them to file and in this way could import them
as posts in a file-based weblog engine like pyblosxom, for example.

[Sam Pullara]

I'd like to see a continuous stream API for new bookmarks similar to
Twitter's Hosebird. This can be used to notice trends and new sites
quickly without as much trouble as a polling API.

Sam

[maciej]

Can you explain how this kind of API would be different from an RSS
(or other) feed of the recent bookmarks page? I can't find any
documentation on Hosebird, but I am glad to see a project with an even
dumber name than FireEagle.

[Sam Pullara]

Thats the internal name, each feed has it own name like firehose,
garden hose and sprtizer:

http://apiwiki.twitter.com/Streaming-API-Documentation

The basic idea is that you connect to a socket and get continuous
updates without additional connections. Amusingly, FireEagle now has a
streaming API like this one as well. Too bad firehose is already
taken :)

Sam

[maciej]

Pinboard has a peak rate of maybe two public bookmarks a minute - more
of a pneumatic tube model of bookmark delivery, rather than any kind
of stream. Given that, a recent items RSS feed seems like the
prudent next step. Clients will have to poll at least once every 50
minutes or risk losing valuable data...

[TylerWeir]

On Aug 3, 2:06 am, maciej <mceglow...@gmail.com> wrote:
> Pinboard has a peak rate of maybe two public bookmarks a minute - more
> of a pneumatic tube model of bookmark delivery, rather than any kind
> of stream.

The Carrier Pigeon of streaming services. :)

[Jarin Udom]

It would be really awesome to have plists as one of the API formats,
since regular XML and JSON parsing can be kind of slow on the iPhone
sometimes and you get plist-to-object mapping for "free".

http://developer.apple.com/documentation/Darwin/Reference/ManPages/man5/plist.5.html

Jarin Udom
Robot Mode LLC

[incanus]

On Jul 26, 7:00 pm, maciej <mceglow...@gmail.com> wrote:

Honestly, I would like to see your basic bookmarking service API calls
such as getting all bookmarks, getting recent bookmarks, posting a new
bookmark, and some sort of hash or other lightweight way of checking
the freshness of older bookmarks. While something like the ill-fated
Ma.gnolia's Mirrord API (where they mirrored Delicious' API) would be
nice for compatibility reasons, I understand your efforts to start
something fresh and new. However, I think the key to adoption is a
robust API.

As for format, any modern format is ideal, but something like XML and/
or JSON would be standard.

[gerwitz]

I want delicious-ish JSON, merely so I can use Lenni's Overdrive:

http://lenni.info/blog/2009/05/overdrive-deliciouscom-a-fast-browser-plugin-for-deliciouscom/

[maciej]

Heh, I remember building something that sucked all your bookmarks from
del and filtered them in javascript. This looks like a cool
implementation of the new local storage.

What if I just provided a way to suck down all bookmarks in JSON
format, similar to the RSS export? Poke Lenni and see if that would
be enough for him to go on.

On Aug 14, 3:32 am, gerwitz <h...@gerwitz.com> wrote:
> I want delicious-ish JSON, merely so I can use Lenni's Overdrive:
>

> http://lenni.info/blog/2009/05/overdrive-deliciouscom-a-fast-browser-...

[Johannes]

Sorry to dig up this old thread - wanted to reference this because it mentions this:

* Download all bookmarks matching a given search query [via API]

Is that something that is on the roadmap?

Am Samstag, 1. August 2009 21:08:29 UTC+2 schrieb seanh: