Bug 8188 - include a "beacon" so we find out about installs
include a "beacon" so we find out about installs
Status: NEW
Product: OJS
Classification: Unclassified
Component: General
3.0
All All
: P3 normal
Assigned To: PKP Support
: 4148 (view as bug list)
Depends on:
Blocks:
  Show dependency treegraph
 
Reported: 2013-04-05 09:57 PDT by Juan Pablo Alperin
Modified: 2014-10-06 17:08 PDT (History)
4 users (show)

See Also:
Version Reported In:
Also Affects:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Juan Pablo Alperin 2013-04-05 09:57:40 PDT
We should have OJS talk back to PKP and give us information about the install. I don't mind making it configurable (via config file), so people can turn it off, but it should be on by default. 

We should report back to PKP some usage stats, such as: 
- number of journals
- some metric of whether or not they are using the workflow

These could be anonymous somehow. I suggest further thinking on how to make it work.
Comment 1 Alec Smecher 2013-04-05 10:48:00 PDT
As we've often discussed, this has the potential to be evil. I think the best compromise we've come up with is to look at referrals to the PKP version descriptor, which ought to result in an upgrade warning appearing upon JM login. If the metrics you're looking for are available on the journal website, then reaching out to that web server as per the current tracker (though perhaps making use of more machine-friendly tools such as the sitemap) ought to be friendly enough to the userbase; if the metrics aren't currently available then I'd argue we ought not to be collecting them behind the scenes.
Comment 2 Juan Pablo Alperin 2013-04-05 10:49:35 PDT
many OSS projects give the option of giving back anonymous usage stats. Eclipse is one example. Wordpress is another.
Comment 3 Juan Pablo Alperin 2013-04-05 10:50:08 PDT
i'd like to know what plugins are enabled, what options are filled in. It would be an invaluable resource for allocating development priorities.
Comment 4 Alec Smecher 2013-04-05 10:53:20 PDT
The version check would be a good, non-evil source for install URLs; an opt-in for further stats similar to Wordpress is a good idea but needs scoping/design.
Comment 5 Juan Pablo Alperin 2013-04-05 10:56:31 PDT
opt-out
Comment 6 Juan Pablo Alperin 2013-04-05 11:00:57 PDT
also, the version check and the sitemap.xml
Comment 7 Alec Smecher 2013-05-29 15:32:46 PDT
*** Bug 4148 has been marked as a duplicate of this bug. ***
Comment 8 Marc Bria 2013-08-17 07:09:18 PDT
Good proposal.
Posting to follow.
Comment 9 Alec Smecher 2013-09-12 15:02:49 PDT
CrossRef has suggested that we trigger the beacon when an OJS install is created, and again when a journal is enabled ("go-live" if the administrator is using the "Enabled" checkbox for that purpose, though I suspect just keeping the url private is also a common means for this).
Comment 10 stuart yeates 2014-10-05 19:31:08 PDT
I think we can do much better

I think we need three check boxes on 'go live'
(a) add this journal to a publically accessible list of journals on "http://pkp.sfu.ca"
(b) add all journals on this install to a publicly accessible list of journals on "http://pkp.sfu.ca"
(c) request that PKP notify the journal manager of this journal of new OJS moajr versions and security fixes as they are released (~ 1 email per 2 months)

We can then make the publicly accessible list of journals accessible to google scholar, so they're picked up automagically...
Comment 11 Alec Smecher 2014-10-06 08:45:35 PDT
Stuart, Google Scholar already picks up any available journals automatically by detecting the Google Scholar specific metadata tags in article pages during regular crawling activity.
Comment 12 stuart yeates 2014-10-06 16:21:34 PDT
"Google Scholar already picks up any available journals automatically" --- really? That is not our experience, with journals such as http://ojs.victoria.ac.nz/LEW/ being seen by google but not google scholar. Is there something wrong with that site?
Comment 13 Alec Smecher 2014-10-06 17:08:23 PDT
Stuart, unfortunately Google Scholar's crawling process is opaque to us. As far as we've been told, putting Google Scholar metadata tags in place and making sure the site is crawlable should be enough. Beyond that, unfortunately, you'd have to contact Google Scholar for details.