To Anon 1: Glad you do all of this.
It is appalling how many people STILL do not do this.
Daniel: GOOD! Yours is simpler and I can’t imagine anyone disagreeing
(but see comment 5 by Mihai Pop on my blog).
Caveat: I would still like to have people
but there OLD papers online that are not,
and to help people find papers. But given what the comments on my manfesto it may well be that yours is the best we can do now.
@Gasarch I am not aware of any case where ACM or IEEE have asked an author to take something off arXiv.
Beside, we are ACM and IEEE. It is not like they could afford to upset all of us.
D. Eppsteinsays:
Yes.
I generally make sure to upload to arXiv prior to sending in the copyright form, and (when I remember this part) to notify the conference or journal publisher of the arXiv version on the form. That way they can plausibly reject my paper for that reason (though they never do) but they have no excuse for asking me to take down the arXiv version later. (Not that takedowns from arXiv are really possible…)
Of course, when I have co-authors handling the details, they may not handle them the same way (I am still bugging my coauthors to put up a public preprint version of one of my papers a month or two after its acceptance).
Also, I’m pretty sure ACM conferences specifically allow their papers to be put up in arXiv (or as they call it the “ACM Computing Research Repository”) but I don’t know about IEEE.
And what about the copyright agreements that have been signed? For big organizations like IEEE or ACM, authors have to transfer their copyright to the organization, that forbids posting in sites like Arxiv. It is allowed to post the papers in your own website though.
Is it really necessary to post to Arxiv, breaking the agreements, when there are tools like Google Scholar available? For instance, all the papers that are in my web can be easily found in Google Scholar. And I am not breaking any agreement that I previously signed.
When you transfer the copyright you usually cannot post the _printed_ version of the article (as it will appear in journal). But, normally you can submit a preprint to the arXive, just your version, as it was submitted to the journal.
As for the Google Scholar – people are used to check the arXive updates, so your paper will be much more visible there.
I would add to the above aims that publication impact can be increased by sharing other resources related to a publication.
While acknowledging real limitations related to time, copyright, privacy, and other issues, publication impact can be increased by sharing following materials in an accessible online form:
1) Raw data and metadata used in the publication.
2) Digital materials related to the publications (e.g., survey materials; tests; software created; etc.)
3) Code used to generate results (e.g., tables and figures).
Thanks for this very good idea. I agree with Jeremy Anglim to add “code and data share” to the list : we need to let others to reproduce our results, easily.
I do 1, 3 and 4 (and will probably start doing 2 also).
For those who use BibTeX, there a great tool by Martin Monperrus called bibtexbrowser to generate a list of publications automatically for your website that takes care of 1, 3 (sort of) and 4 (the new version has rss feeds)…
Ed Felten has suggested that all scientists start using the same word processors that the professional programmers are using. It has a very precise and sophisticated versoning system and is able to handle much information.
“Developing Texts Like We Develop Software” http://www.freedom-to-tinker.com/blog/felten/developing-texts-we-develop-software
While many researchers are already committed to these guidelines, an explicit manifesto will give added visibility for these objectives and helps to define the essential baseline. Support.
To Anon 1: Glad you do all of this.
It is appalling how many people STILL do not do this.
Daniel: GOOD! Yours is simpler and I can’t imagine anyone disagreeing
(but see comment 5 by Mihai Pop on my blog).
Caveat: I would still like to have people
but there OLD papers online that are not,
and to help people find papers. But given what the comments on my manfesto it may well be that yours is the best we can do now.
@herraiz No, it is not necessary to post to arXiv in what I wrote above.
Have organizations like IEEE or ACM or Elsevier or others actually told authors to take things off of arXiv? I ask this nonrhetoically.
@Gasarch I am not aware of any case where ACM or IEEE have asked an author to take something off arXiv.
Beside, we are ACM and IEEE. It is not like they could afford to upset all of us.
Yes.
I generally make sure to upload to arXiv prior to sending in the copyright form, and (when I remember this part) to notify the conference or journal publisher of the arXiv version on the form. That way they can plausibly reject my paper for that reason (though they never do) but they have no excuse for asking me to take down the arXiv version later. (Not that takedowns from arXiv are really possible…)
Of course, when I have co-authors handling the details, they may not handle them the same way (I am still bugging my coauthors to put up a public preprint version of one of my papers a month or two after its acceptance).
Also, I’m pretty sure ACM conferences specifically allow their papers to be put up in arXiv (or as they call it the “ACM Computing Research Repository”) but I don’t know about IEEE.
Hm. I do all of the following. I even post my conference posters online.
*all of the above, sorry 🙂
And what about the copyright agreements that have been signed? For big organizations like IEEE or ACM, authors have to transfer their copyright to the organization, that forbids posting in sites like Arxiv. It is allowed to post the papers in your own website though.
Is it really necessary to post to Arxiv, breaking the agreements, when there are tools like Google Scholar available? For instance, all the papers that are in my web can be easily found in Google Scholar. And I am not breaking any agreement that I previously signed.
2 herraiz:
When you transfer the copyright you usually cannot post the _printed_ version of the article (as it will appear in journal). But, normally you can submit a preprint to the arXive, just your version, as it was submitted to the journal.
As for the Google Scholar – people are used to check the arXive updates, so your paper will be much more visible there.
I love it! I’ll make sure to add explicit (and simple) support for this idea.
Cheers,
christopher
I would add to the above aims that publication impact can be increased by sharing other resources related to a publication.
While acknowledging real limitations related to time, copyright, privacy, and other issues, publication impact can be increased by sharing following materials in an accessible online form:
1) Raw data and metadata used in the publication.
2) Digital materials related to the publications (e.g., survey materials; tests; software created; etc.)
3) Code used to generate results (e.g., tables and figures).
Additional Thoughts
My own post on the topic:
http://jeromyanglim.blogspot.com/2009/09/linking-text-results-and-analyses.html
A discussion on Andrew Gelman’s blog:
http://www.stat.columbia.edu/~cook/movabletype/archives/2009/09/we_received_onl.html
thanks, Daniel,
for a collaborative writing project on this issue see here:
http://etherpad.com/An-Open-Publishing-Manifesto
feel free to join in everyone 🙂
Totally!
Actually, it is very nice to see that nowadays in machine learning, most of the decent researchers actually do all these.
Thanks for this very good idea. I agree with Jeremy Anglim to add “code and data share” to the list : we need to let others to reproduce our results, easily.
I do 1, 3 and 4 (and will probably start doing 2 also).
For those who use BibTeX, there a great tool by Martin Monperrus called bibtexbrowser to generate a list of publications automatically for your website that takes care of 1, 3 (sort of) and 4 (the new version has rss feeds)…
I blogged about it here: http://jorgecardoso.eu/blog/index.php?/archives/154-Bibtexbrowser.html
The tool can be found here: http://www.monperrus.net/martin/bibtexbrowser/
Ed Felten has suggested that all scientists start using the same word processors that the professional programmers are using. It has a very precise and sophisticated versoning system and is able to handle much information.
“Developing Texts Like We Develop Software”
http://www.freedom-to-tinker.com/blog/felten/developing-texts-we-develop-software
While many researchers are already committed to these guidelines, an explicit manifesto will give added visibility for these objectives and helps to define the essential baseline. Support.
Claudia: the link http://etherpad.com/An-Open-Publishing-Manifesto is broken?
Note (in general): internet links & sites become obsolete.
Etherpad was acquired by Google, and then open-sourced. See: http://etherpad.com/ for links to the source repository, and other info.