arXiv trackbacks

[fgdorais]

It makes sense to send trackbacks to the arXiv in the same way that MathOverflow does.

[cjlee112]

good idea, worth learning about further.

[fgdorais]

Talk to people at the arXiv. MathOverflow has an "easy-pass" for their editorial mechanism. They may consider doing the same for spnet.

[johncarlosbaez]

Here's a complementary issue. Would it be possible to take all existing arXiv trackbacks and automatically turn them into entries on the spnetwork? If so, this might be a quick way to get information on a lot arXiv articles into the spnetwork database.

[cjlee112]

Yeah, I immediately thought that when I first read this issue. That's what I want to research. A couple thoughts:

• right now, at least, this does not have much traffic, maybe 3 or 4 new links per day. I don't know what the total number of all links they've collected is, or whether they'd be willing to share them with us.
• they've configured this purely for users clicking on their site. That is, the only way to get the trackbacks is to go to http://arxiv.org/tb/recent and start clicking the links to see what you get redirected to. In the absence of arXiv supplying an API or explicitly telling us we can scrape those links, we probably shouldn't. arXiv does not like robots.

[fgdorais]

I don't think it's "appropriate" to "steal" trackbacks. A great deal of trackbacks come from MathOverflow and it wouldn't be too hard to get them from the source. Let us know how we can help.

[fgdorais]

The arXiv trackbacks are available by clicking the "Blog Links" on the paper's abstract page http://arxiv.org/abs/1305.6158, or directly via the url scheme http://arxiv.org/tb/1305.6158.

In December 2012, MathOverflow had generated trackbacks to over 5000 different arXiv articles.

[ribault]

It is crucial to systematically have trackbacks on arXiv. I am considering using spnetwork, but in my community (hep-th) nobody has heard of it. Trackbacks on arXiv should bring the network to the attention of researchers, without the need for active proselytizing. If there are no trackbacks people (including me) will be reluctant to write comments in spnetwork, for fear such comments will be lost to the community.

[johncarlosbaez]

In most communities nobody has heard of the spnetwork, and spreading the word will be a continuing challenge. I urge everyone who is not a mathematician to blog about it and spread the word.

However, I agree that we should get the arXiv to have trackbacks to the spnetwork - this would be a great way to get more people to notice it and use it. I will email some arXiv people about that.

[cjlee112]

@fgdorais OK, who at arXiv should we talk to about getting our trackbacks passed in expedited form like MathOverflow? And who at MathOverflow should we speak to about getting MathOverflow's arXiv trackbacks in order to index them and link to the appropriate content on MathOverflow?

[fgdorais]

@johncarlosbaez said elsewhere he was going to talk to people at the arXiv. Let me know by email if that doesn't pan out.

We recently upgraded our software which unfortunately broke our trackback mechanism. We'll fix it and start sending them to you along at the same time.

[kim-em]

I'm more or less the one responsible for MathOverflow generating trackback. I am planning on fixing it, although it makes sense to wait for a few weeks as the Stack Exchange API isn't active yet for MathOverflow, and clearly this is the right way to do things.

When it gets fixed, we will very happily send trackbacks towards you guys too!

[cjlee112]

@semorrison forgive my ignorance of your system; "send trackbacks toward" us means you'd POST json data for each trackback to some REST endpoint URL on our server?

[kim-em]

The usual trackback mechanism is simply to make a GET request with a few parameters. Here's an example of how we notify the arXiv:

 http://arxiv.org/trackback/1307.1702?blog_name=MathOverflow&title=A%20Question&url=http://mathoverflow.net/questions/123

Obviously we can do any similarly simple thing! I'm sure we can go so far as to prepare some JSON data for you, if that would make you happy :-)

[cjlee112]

I assumed POST mainly because it's the RESTful verb for adding data to a collection. I was just trying to understand your usual mechanism.

[kim-em]

I presume that POSTing would work just as well (and I appreciate that it is more correct). I was just reporting what our existing (now broken) script at MathOverflow does.

On Tue, Jul 9, 2013 at 2:49 PM, cjlee112 notifications@github.com wrote:

I assumed POST mainly because it's the RESTful verb for adding data to a collection. I was just trying to understand your usual mechanism.

— Reply to this email directly or view it on GitHubhttps://github.com/cjlee112/spnet/issues/27#issuecomment-20653313 .

[kim-em]

I've asked the arXiv administrators to approve trackbacks #spnetwork sends to them.

[kim-em]

They reply:

Dear Scott Morrison,

We apologize for the delay in our reply. If spnetwork generates trackbacks they'll be ingested normally.

We have experimental set of data listing existing trackbacks at:

http://resync.library.cornell.edu/lod/tb/all

This currently has only the source and destination of each trackback expressed using the Open Annotations data model. We plan to add other details but cannot guarantee a schedule, or whether the format might change.

arXiv admin