The intext citations work well and by changing preferences can "back date" older documents to solve this problem. I note however that in the reference (bibliography), some titles are all caps and others in upper and lower case. Looking back at the core reference, it seems that some journals put all titles in all CAPs and Zotero downloads them faithfully. Is there an equivalent fix to harmonise all bibliographic outputs as well? This one is less critical as it affects far fewer references. But it would be nice if the citation routines could harmonise all the output almost irrespective of how Zotero downloads and stores the "raw" citation data.
The problem or really a 'bug' is not that Zotero attempts to disambiguate between John M. Smith and J. M. Smith (it indeed should do this), but that it disambiguates in-text citations based on what is in the *entire* database, not what is in the bibliography.
If you have a database with John M. Smith, Fred Smith, and Horace Smith, it seems like you will always get first initials or names even if your bibliography only contains entries from John M. Smith. I have gone through and cleaned up all entries for a particular authors so that all are identical. But you can't 'clean up' the entries for 2 different authors with the same last name in your database--even if you only use 1 of them in a paper.
I just ran a simple test on a fresh document with ersatz entries, each with the same last name, and with differing first names, in the Chicago Author-Date style. I get the author's name alone when citing only one work. When I add the other, I get the full name of both authors. When I delete the first entry and refresh citations, it reverts to the last name alone. It appears that everything is working as advertised.
Do you have a specific combination of data and style that produces a different result? The problem needs to be reproducible before we can properly call it a bug.
Bug, feature, enhancement.... a bit of each. Perhaps a future release might offer a very intuitive means of creating preferences for ambiguation or even a pop up when generating a bibliography asking how one wants to handle similar names. I use Harvard style and I have two choices - one says something about disambiguation and that is the one I use which seems to help a lot.
However we resolve it, I guess the advice I give my students holds. Don't produce a thesis, press the button and just hand it in assuming the software has done it all correctly. One must still go through the citations and references to make sure all is as it should be!!
Hmmm. I'm glad that it is working correctly in in some circumstances. I'll run some controlled tests to see if I can isolate situations in which it is working incorrectly.
The test I did is with a paper I'm doing now with a lot of citations. I have a number of citations by me (C. Michael Barton) in the paper. I've gone through all of my pubs in my database and I have my name entered in exactly the same way (I use the autocomplete feature to ensure this). My database also has entries by a couple of other Bartons, but neither of them is cited in this particular paper. I've also made sure that each other Barton is in the database in a single way only. Nonetheless, all the citations in the paper are coming up C. M. Barton.
One thing I note is that the autocomplete maintains a variety of other ways of listing my name (e.g., Barton, C M; Barton, C. M.; Barton, C. Michael; Barton, C. Michael.) from various imported formats that I've since edited for consistency. Is there any way to clear the autocomplete cache in order to test if this is a problem here?
hmmm that's weird and would certainly be a bug - we need a way to replicate this, though.
Could you maybe export all your C.M. Barton citations as Zotero RDF (without note and attachments) and either upload the rdf file somewhere or paste it to
gist.github.com (just paste in the window and "create public gist).
About auto-complete - not sure, have you tried emptying your trash?
Just a quick query, but does the cited paper have a second author with the same surname? I have noticed the behavior you describe when this is the case.
I did some tests and now understand the situation. The disambiguating is indeed not working quite right, but it's closer to right than I thought.
After exporting records, trying them in dummy paragraphs, and doing more digging, I found that I DID have another Barton in a cited reference. However, this Barton (R.N.E. Barton) was the 5th author in a 6 author citation that was inserted as (d'Errico et al 2009). This is why I couldn't find it in any straightforward manner. So the disambiguating by adding my initials makes sense now, even though it is not necessary.
To refine the disambiguating, it would be good if it could check to see if the potential ambiguous citation is further down in a multiple author sequence than the limit specified by the code for invoking 'et al' (I don't remember the XML syntax, but have seen it).
A second thing that I noticed is that the reference that gets the initials added to disambiguate seems to be the first one in the database (C. Michael Barton comes before R.N.E. Barton). But it might be better to be the 2nd (and 3rd, etc) cited instead.
Finally, a simple check button to 'turn off disambiguating initials' added to the citation entry dialog would make fixing this manually quick and easy (as opposed to opening the editor) if the automatic enhancements are too difficult to do. For multiple citations, the check box could be next to each reference in the list. Even better would be the abilit to turn it off once for a reference and have it stay turned off (or on) for that reference any other time it is cited.
Related to this I have a question that could speed up the cleaning of databases.
I opened up the Zotero.sql database using SQLite Manager for FireFox. In the createrData table, I can see the entries for the author autocomplete, including the ones that I no longer use or want to use. What happens is I edit this table by...
1) putting the same first name version in all instances of an author with different versions of the first name. That is, replacing John M Smith, J M Smith, J. M. Smith, and JM. Smith with John M Smith in all cases?
2) deleting the unused entries?
3) deleting the incorrect entries whether they are used or not?
OK thanks for the update - I was aware of this issue and I believe so is
Frank "disambiguation" Bennett.
For the second thing - all authors with the same last name get disambiguated by adding initials and first name- and I think that's standard practice in citation styles, too.
Finally - yeah, the checkbox might be an option - unfortunately my impression is that every time we hit something like this the first suggestion is to put an ad hoc solution in the plugin. Considering that there is a bunch of issues where automation is a little tricky with citations, I'm quite concerned about a completely bloated/overloaded plugin. So any way we can solve this the right way would be great.
What adamsmith says (what I wouldn't do to lose that nickname, though!).
I did seem to remember that hidden names can trigger disambiguation. Then I did a test earlier today, and they didn't seem to -- but I probably messed up in setting up the test. Disambiguation can be confusing, especially when it happens unexpectedly. Nothing more frustrating, at the end of an afternoon spent fighting with a hopelessly ill-drafted, incomprehensible bug-infested useless excuse for a computer program, than to discover that it has been doing exactly what you instructed it to do.
As a note, though - Zotero 2.1 and later with a fresh install of the APA style from the repository use disambiguation _exactly_ as prescribed by the APA style guide. So if you're removing the option you're likely not complying with APA anymore.
This won't get "fixed" because there is nothing to fix. Zotero needs to be able to disambiguate authors - and styles require that to be done.
You can find information on removing disambiguation from a style here:
http://www.zotero.org/support/dev/citation_styles/style_editing_step-by-step#examples_edits
I managed to fix "my" style with all the help available on Zotero forum. Also thank you adamsmith for your quick reply. I also found out that sometimes a restart of Firefox(Zotero) and my word editor(LibreOffice) does magic to citations. Some things simply change for the better for no apparent reason. :)
Well any way I fixed may issues with all the advice that was given above. But it would be nice to have some kind of an editor to edit the styles. I mean a more graphical way of doing it. But that is probably an entirely new topic....
Upgrade Storage
The intext citations work well and by changing preferences can "back date" older documents to solve this problem. I note however that in the reference (bibliography), some titles are all caps and others in upper and lower case. Looking back at the core reference, it seems that some journals put all titles in all CAPs and Zotero downloads them faithfully. Is there an equivalent fix to harmonise all bibliographic outputs as well? This one is less critical as it affects far fewer references. But it would be nice if the citation routines could harmonise all the output almost irrespective of how Zotero downloads and stores the "raw" citation data.
As always, many thanks
If you have a database with John M. Smith, Fred Smith, and Horace Smith, it seems like you will always get first initials or names even if your bibliography only contains entries from John M. Smith. I have gone through and cleaned up all entries for a particular authors so that all are identical. But you can't 'clean up' the entries for 2 different authors with the same last name in your database--even if you only use 1 of them in a paper.
This is a bug AFAIC.
I just ran a simple test on a fresh document with ersatz entries, each with the same last name, and with differing first names, in the Chicago Author-Date style. I get the author's name alone when citing only one work. When I add the other, I get the full name of both authors. When I delete the first entry and refresh citations, it reverts to the last name alone. It appears that everything is working as advertised.
Do you have a specific combination of data and style that produces a different result? The problem needs to be reproducible before we can properly call it a bug.
However we resolve it, I guess the advice I give my students holds. Don't produce a thesis, press the button and just hand it in assuming the software has done it all correctly. One must still go through the citations and references to make sure all is as it should be!!
Hmmm. I'm glad that it is working correctly in in some circumstances. I'll run some controlled tests to see if I can isolate situations in which it is working incorrectly.
The test I did is with a paper I'm doing now with a lot of citations. I have a number of citations by me (C. Michael Barton) in the paper. I've gone through all of my pubs in my database and I have my name entered in exactly the same way (I use the autocomplete feature to ensure this). My database also has entries by a couple of other Bartons, but neither of them is cited in this particular paper. I've also made sure that each other Barton is in the database in a single way only. Nonetheless, all the citations in the paper are coming up C. M. Barton.
One thing I note is that the autocomplete maintains a variety of other ways of listing my name (e.g., Barton, C M; Barton, C. M.; Barton, C. Michael; Barton, C. Michael.) from various imported formats that I've since edited for consistency. Is there any way to clear the autocomplete cache in order to test if this is a problem here?
Could you maybe export all your C.M. Barton citations as Zotero RDF (without note and attachments) and either upload the rdf file somewhere or paste it to
gist.github.com (just paste in the window and "create public gist).
About auto-complete - not sure, have you tried emptying your trash?
Just a quick query, but does the cited paper have a second author with the same surname? I have noticed the behavior you describe when this is the case.
I'll also create a mini-bilbiography that I can test further and post for Adam and others to look at.
After exporting records, trying them in dummy paragraphs, and doing more digging, I found that I DID have another Barton in a cited reference. However, this Barton (R.N.E. Barton) was the 5th author in a 6 author citation that was inserted as (d'Errico et al 2009). This is why I couldn't find it in any straightforward manner. So the disambiguating by adding my initials makes sense now, even though it is not necessary.
To refine the disambiguating, it would be good if it could check to see if the potential ambiguous citation is further down in a multiple author sequence than the limit specified by the code for invoking 'et al' (I don't remember the XML syntax, but have seen it).
A second thing that I noticed is that the reference that gets the initials added to disambiguate seems to be the first one in the database (C. Michael Barton comes before R.N.E. Barton). But it might be better to be the 2nd (and 3rd, etc) cited instead.
Finally, a simple check button to 'turn off disambiguating initials' added to the citation entry dialog would make fixing this manually quick and easy (as opposed to opening the editor) if the automatic enhancements are too difficult to do. For multiple citations, the check box could be next to each reference in the list. Even better would be the abilit to turn it off once for a reference and have it stay turned off (or on) for that reference any other time it is cited.
I opened up the Zotero.sql database using SQLite Manager for FireFox. In the createrData table, I can see the entries for the author autocomplete, including the ones that I no longer use or want to use. What happens is I edit this table by...
1) putting the same first name version in all instances of an author with different versions of the first name. That is, replacing John M Smith, J M Smith, J. M. Smith, and JM. Smith with John M Smith in all cases?
2) deleting the unused entries?
3) deleting the incorrect entries whether they are used or not?
Thanks
Frank "disambiguation" Bennett.
For the second thing - all authors with the same last name get disambiguated by adding initials and first name- and I think that's standard practice in citation styles, too.
Finally - yeah, the checkbox might be an option - unfortunately my impression is that every time we hit something like this the first suggestion is to put an ad hoc solution in the plugin. Considering that there is a bunch of issues where automation is a little tricky with citations, I'm quite concerned about a completely bloated/overloaded plugin. So any way we can solve this the right way would be great.
I did seem to remember that hidden names can trigger disambiguation. Then I did a test earlier today, and they didn't seem to -- but I probably messed up in setting up the test. Disambiguation can be confusing, especially when it happens unexpectedly. Nothing more frustrating, at the end of an afternoon spent fighting with a hopelessly ill-drafted, incomprehensible bug-infested useless excuse for a computer program, than to discover that it has been doing exactly what you instructed it to do.
Or did I miss something in the above conversation?
Thanks!
You can find information on removing disambiguation from a style here:
http://www.zotero.org/support/dev/citation_styles/style_editing_step-by-step#examples_edits
Well any way I fixed may issues with all the advice that was given above. But it would be nice to have some kind of an editor to edit the styles. I mean a more graphical way of doing it. But that is probably an entirely new topic....