submit to arXiv?

[gusl]

Should I hesitate to submit my paper to arXiv? I didn't give up any rights when I gave UAI permission to publish it in their proceedings.

Comments

[simrob.livejournal.com]

I say do it.

[bhudson.livejournal.com]

Always. If people can't download your work, it doesn't exist.

[bhudson.livejournal.com]

Oh wait -- actually, it may make more sense for this to be copied into an SCS tech report.

[gustavolacerda.livejournal.com]

but it can be downloaded from my web page!

[bhudson.livejournal.com]

then it doesn't matter so much

[gustavolacerda.livejournal.com]

well, I want more people to see it.

OTOH, I'd like to avoid duplication. Right now it's easy to keep my version up-to-date, but if arXiv keeps a static version, this will no longer be the case.

[bhudson.livejournal.com]

People don't go to arxiv to find CS papers (yet); your web page is as good as anywhere else.

But doesn't arxiv have a revision history thing?

[simrob.livejournal.com]

Yes, there's version control, but I don't think it's a bad thing to have old copies of your papers floating around, I think it's a good thing as long as there is good version control. It allows people to keep track of the paper they downloaded as a web page, and that web page will inform them if their copy is out of date.

This isn't hypothetical, *yesterday* I found that a paper my reading group had read had a new version simply by following the link to arXiv in the old reading group email.

[bhudson.livejournal.com]

I was responding to "arxiv keeps a static version" which is almost perfectly wrong. Although I suppose it does not keep a version where the equations sing and dance for you.