IngentaConnect, unable to add

zpinhead

it does not seem as if the page is recognized.
here is one example, but i can't find any pages that work.

http://www.ingentaconnect.com/content/routledg/cpes/2005/00000010/00000003/art00006

when i add the page, it adds as a web page -- the web address is saved is:
http://an.tacoda.net/an/tpp.html and the title of the page is 'TPP'


and the 'label' visible is 'Untitled', as if the url passed to detectWeb is not
not that of the actual page, but of some advertiser, or akamai cache?
i think this is more a matter of being recognized properly, than of the actual extraction not working.
firefox 3.5.3 -- Windows XP.
zotero 2.0b7.2

zpinhead

ingentaconnect has had a lot of trouble, there are previous posts from 2008,
i believe, and 2007, where problems were solved;

Bionatsci

The link you provided adds successfully as a journal article for me (zotero 2.0b7.2, firefox 3.5.3 - Windows Vista), although the quality of the scraped data is not great - e.g. nothing gets put in the issue field, the issue number is put into the date field and the author is recorded as "Pope, Clive [1]".

when i add the page, it adds as a web page -- the web address is saved is:
http://an.tacoda.net/an/tpp.html and the title of the page is 'TPP'

The URL saved is correct for me, are you by any chance connecting to IngentaConnect through a university proxy? AFAICR there have been some reports of proxied connections causing bugs in certain translators in the past and I guess that could be the issue here.

dstillman

the quality of the scraped data is not great - e.g. nothing gets put in the issue field, the issue number is put into the date field and the author is recorded as "Pope, Clive [1]"

ER  - 
TY  - JOUR
AU  - Pope, Clive [1]
TI  - Once more with feeling: affect and playing with the TGfU model
JO  - Physical Education and Sport Pedagogy
PY  - ///Number 3/November 2005
VL  - 10
SP  - 271
EP  - 286
UR  - http://www.ingentaconnect.com/content/routledg/cpes/2005/00000010/00000003/art00006
M3  - doi:10.1080/17408980500340885
UR  - http://dx.doi.org/10.1080/17408980500340885

As you can see, they're putting issue into the date field, and the "[1]" is in the author field. (They also put "ER", which is supposed to be the last field of a RIS entry, at the beginning, but Zotero's translator adjusts for that.) If the issue/date and author problems occur on many records on the site, we could probably adjust for them, but these are clearly their bugs.

zpinhead: Your problem, clearly, is different. Generate debug output from the Advanced pane of the Zotero prefs for loading the page and attempting to save, submit the output to the server, and post the Debug ID here. Thanks.

zpinhead

I will try enabling the debug, will need to do this tonight.
@Bionatsci this is not through a proxy. it's not even recognized by the detectWeb --
there is something screwy in what is passed into detectWeb.

zpinhead

i enabled the debug logging, and visited and saved an ingenta reference:
somehow this tacoda.net, some sort of advertising site, is interposing itself.


(4)(+0000000): Translate: Binding sandbox to http://an.tacoda.net/an/tpp.html

(3)(+0000000): Translate: Searching for translators for http://an.tacoda.net/an/tpp.html

(4)(+0000002): Translate: Parsing code for Untitled

(4)(+0000001): Translate: Parsing code for unAPI

(4)(+0000003): Translate: Executed detectCode for unAPI

(4)(+0000000): Translate: Parsing code for COinS

(4)(+0000003): Translate: Executed detectCode for COinS

(4)(+0000000): Translate: Parsing code for DOI

(4)(+0000002): Translate: Executed detectCode for DOI

(4)(+0000000): Translate: Parsing code for Embedded RDF

(4)(+0000003): Translate: Executed detectCode for Embedded RDF

(5)(+0000000): Translate: running handler 0 for translators

(4)(+0004930): Translate: Parsing code for Untitled

(4)(+0000024): Translate: Binding sandbox to http://www.example.com/

(4)(+0000001): Translate: Parsing code for RDF

(4)(+0000054): Translate: Setting configure option dataMode to rdf

dstillman

That doesn't look like the whole debug output. Can you clear the output, load the page, submit the output, and provide the Debug ID?

zpinhead

Dan,

I just tried loading the addon 'AdBlocker Plus'
https://addons.mozilla.org/en-US/firefox/addon/1865
and selecting the default list of add sites to block. when that is done, the
ingenta connect detect web works properly!

so i disabled the add on again, and that recreates the error.

this only represents visiting the page, not saving.
I am figuring that's where we went wrong -- as the fact that this is an ingenta page is
not even discovered because the page url is reported as http://an.tacoda.net/an/tpp.html


D196344970, debugid.
(4)(+0000000): Translate: Binding sandbox to http://an.tacoda.net/an/tpp.html

(3)(+0000001): Translate: Searching for translators for http://an.tacoda.net/an/tpp.html

(4)(+0000001): Translate: Parsing code for Untitled

(4)(+0000001): Translate: Parsing code for unAPI

(4)(+0000003): Translate: Executed detectCode for unAPI

(4)(+0000000): Translate: Parsing code for COinS

(4)(+0000002): Translate: Executed detectCode for COinS

(4)(+0000000): Translate: Parsing code for DOI

(4)(+0000002): Translate: Executed detectCode for DOI

(4)(+0000000): Translate: Parsing code for Embedded RDF

(4)(+0000002): Translate: Executed detectCode for Embedded RDF

(5)(+0000000): Translate: running handler 0 for translators

zpinhead

sorry, i should have included this url:

http://www.ingentaconnect.com/content/bsc/ejb/2007/00000274/00000011/art00022
but i have not been able to save references from any ingenta connect page that i have visited so far.

dstillman

Can you reproduce this with all other Firefox extensions disabled?

What does it says in the tooltip when you hover over the address bar icon?

zpinhead

I disable all extensions, except zotero -- and this still happens.
if i re-enable the 'Adblocker Plus' extension then I can indeed save the reference.

the tooltip says: Zotero(Untitled) when the adblocker is DISABLED.
the tooltip says: Zotero(IngentaConnect) when the adblocker is ENABLED.

in all versions of windows vista that i have tried, the z does work.
ingenta explains that they are putting advertising on SOME pages:
http://eyetoeye.ingenta.com/library/issue24/news-advertising.htm
with the "help" of tacoda.net.

dstillman

Have you by any chance been trying to create custom translators? "Untitled" is clearly not a valid translator. (I see now that you noted this above, though I didn't realize at the time what you were referring to.)

Rename the 'translators' directory within your Zotero data directory, and then restart Firefox.

zpinhead

Dan, thanks for your help -- i still do not know what was going on, but
renaming the existing translators directory, and copying translators from a
clean profile of firefox cleared up the problem! It's true that I have been working
on "small" changes in the ISI translator" the ISI translator does not work
for all of its databases -- how this infected the ingenta connector -- i cannot fathom.

I still do not understand why using 'ad blocker' would also solve the problem(??).

anyway, thanks!
Rick

dstillman

It had nothing to do with Ingenta. You had somehow created an invalid translator ("Untitled") that was being triggered on the URL from the ad before it even got to the Ingenta URL.

dstillman

Also, there was no need to copy the translators from another installation—Zotero creates that directory on startup if it doesn't exist.

migugg

I encounter similar problems, debug ID is:
D377453418
for example when I go to:
http://www.ingentaconnect.com/content/beech/spp/2000/00000027/00000003;jsessionid=5ptom9fshcebb.alice
and I click the zotero folder, it saves an item as webpage with the title TPP and then url: http://an.tacoda.net/an/tpp.html
the same happens if I click to a specific article and try to zotero.
I do not run adblocker and I do not have my own translator installed.