estimate of numbers of CSL users?

Related to an off-list chat on value of hacking in the academy got me
wondering: could we come up with a rough estimate of the number of CSL
users (e.g. the sum of all users of CSL implementations)?

Bruce

It’s going to be hard, since I think Papers2 would be the only product
with somewhat reliable user numbers.
We have

  • Zotero
  • Mendeley
  • Papers2
  • Qiqqa
  • Colwiz
  • various citeproc-hs etc. implementations

For Zotero, Sean said that last May
http://quintessenceofham.org/2011/05/21/on-usage-figures/#comments
there were >4million downloads, ~620k accounts, and ~275k daily sync
instances. That’s likely gone up since. My guess would be that the
actual # of users is somewhere between the two later figures, probably
closer to the 275k.

I know Zotero increased it’s storage accounts by 115% in 2011, though
I don’t know how that matches broader numbers.

In any case, I think for those of us with academic jobs, it would be
good to have a number we can put on activity reports, etc. If others
would like to let me know their user estimates, please post here, or
send me off-list.

I also think it would be nice to find ways, for the future
sustainability of this project, to get credit for the intellectual and
other contributions we make (though am not sure how).

Bruce

Don’t forget about Drupal/Biblio http://drupal.org/project/biblio ~3500
web sites, but it would be hard to quantify “users” from that.

Ron.

citeproc-ruby is manly ‘used’ by other projects, so here, too, it’s difficult to say how many actual users there are. There is a wiki list of projects using bibtex-ruby (*) many of which will use citeproc-ruby as well. In addition, I’ve heard from the Berkman Center and a number of university departments; mostly citeproc-ruby seems to be used to format citations on websites.

(*) https://github.com/inukshuk/bibtex-ruby/wiki/Projects-Using-BibTeX-Ruby

Finally, you can look at the ruby gem download statistics here:

(**) https://rubygems.org/profiles/inukshuk

Sylvester

signature.asc (163 Bytes)

Related to an off-list chat on value of hacking in the academy got me
wondering: could we come up with a rough estimate of the number of CSL
users (e.g. the sum of all users of CSL implementations)?

Bruce

I’ve been tracking downloads of MLZ by region, filtering out the
obvious robots. At last count, the client had been downloaded to 1,988
unique IPs across 60 countries (not counting “Europe”, “Asia”, and
"Unknown"). As an experimental project it’s not surprising that the
download count is very small, but the breadth of country coverage has
been good to see.

Frank

There is a PLoS blog post from July last year which stated “The
Mendeley software has been downloaded one million times” [1]. The
vast majority of those will have been versions of Mendeley that
support CSL 1.0 using citeproc-js and using styles from the CSL
repository.

Regards,
Rob.

[1] http://blogs.plos.org/mfenner/2011/07/27/mendeley-1-0-released-today/

Hi guys,

We understand the goal of having an estimate of CSL users, but at the moment, I am afraid we can’t share numbers…

charles

No problem Charles.

In an ideal world CSL users wouldn’t know that they are CSL users :wink:

Absolutely.

Hopefully a few more users of CSL with the Papers for Windows release: http://mekentosj.com/papers/win

Since it’s a complete rewrite with a native Windows implementation, you can count that as one more CSL app :slight_smile:

charles

The CSL processor also has been rewritten?

RintzeOn Tue, Feb 14, 2012 at 1:13 PM, Charles Parnot <@Charles_Parnot>wrote:

Good question, and the answer is yes…

Hello,

Good question, and the answer is yes…

Hmm. The all-in-one-js-file version of citeproc-js
we currently ship is >11K lines of code (with no comments). Unless
there is a lot of code that is not required for typical use cases or
code that can be re-implemented more succinctly, that sounds like a
lot to reproduce faithfully.

Regards,
Rob.