Multilinguality (Zotero vs MLZ vs Juris-M)

I have never quite understood why Zotero didn't (and doesn't) integrate the great multilingual features that Frank Bennett put into MLZ. I am talking here about the ability to have separate fields for the title (and author) in the original language, one for the transliteration of that title/author (which presumably could at some point also be done on the fly), and one for the translation of the title in whatever language one is writing in. It seems to me that it's not because most English authors barely cite any non-English literature, that those of us who do should be penalized. Hence my questions:

  • will Zotero ever include this (increasingly important) functionality? Put otherwise: what prevents Zotero from including these features?

  • (to Frank) - is MLZ now totally replaced by Juris-M? I ask because I see that the MLZ-site is still up - maybe you could provide a link there to Juris-M with some explanation? The new name now suggests that it's mostly for lawyers as opposed to (as I thought) just Zotero with some enhanced functionalities for both multilingual citations AND for laywers

  • does Juris-M have all the functionalities of Zotero?

  • as Zotero 5.0 rolls out, will any innovations in the Zotero code also be reflected in Juris-M? 

  • can we use the same data directory (and syncing options) without fear of messing things up? Also looking forward? 

Thanks (and I really still hope that these two projects will (re-)join, because this is quite confusing to most of us)
  • what prevents Zotero from including these features?
    many thousands of lines of code that would need to be reviewed and merged. In addition, in order to keep sync compatible with Zotero, Frank uses a number of workarounds that wouldn't be suitable for the core code. And especially given the significant impact this has on the structure of the database, all of this isn't easy. I think generally the hope is still to merge this, but I wouldn't get my hopes up for the next couple of years.
    is MLZ now totally replaced by Juris-M?
    Yes. Frank wanted a clearer separation in name between Zotero and Juris-M to avoid confusion.
    does Juris-M have all the functionalities of Zotero?
    Yes, unless you count support from Dan & Adomas as Zotero functionality ;).
    as Zotero 5.0 rolls out, will any innovations in the Zotero code also be reflected in Juris-M?
    Frank is working on getting Juris-M over to 5.0 and I think it's looking pretty good. When that's successful, the answer is yes.
    can we use the same data directory (and syncing options) without fear of messing things up? Also looking forward?
    No on the data repository. That hasn't been possible for quite some time. Yes on the sync, though, and since Frank won't run his own sync server, that's going to stay that way.
  • edited April 26, 2017
    Juris-M replaces MLZ. I'll add a note to clarify at the MLZ link. Thanks for mentioning that.

    Juris-M is happily cruising along as a shadow version of Zotero for the present. The 4.0/5.0 watershed was a scary prospect when the Zotero 5.0 rewrite was first announced, but we've cleared that hurdle, thanks in part to the excellent work Zotero have put into the new automated test suite for their client. I've extended that a little in places to test Juris-M features, and it made the porting work both faster than expected, and more certain in its results.

    We have Juris-M 5.0 Standalone running in the lab now, and a beta release via the Juris-M website will go out after Zotero 5.0 is officially released. Locally, Juris-M has become an important support tool for research and instruction, so I have strong incentives to keep it in good nick.

    As @adamsmith says, Juris-M is sync-compatible with Zotero: a Zotero and a Juris-M client should be able to sync the same group library. When Juris-M is installed on a system that has a pre-existing Zotero library, it will make a copy of the data and convert the schema in the copy (in a database named "jurism.sqlite"). The original data ("zotero.sqlite") will remain in place. You wouldn't want to run Juris-M and Zotero in parallel (I'm not sure whether the word processor plugins would be happy with that), but running them separately should not be a problem.

    I can't speak for the Zotero development team, but as @adamsmith says, there is a lot of code involved. There is also a serious maintenance issue. Juris-M depends on two large data sets: the IANA Subtag Registry (for languages) and a homespun Legal Resource Registry (LRR) (for legal jurisdictions). The former is basically plug-and-play, but the latter is a bespoke resource that wants significant support - ontology design, liaison with local courts and legislative bodies, infrastructure coding. The LRR enables jurisdiction-specific citation style modules (so that UK, US, and Japanese legal references can be presented correctly in the same document, for e.g.), and the work of coordinating the drafting and maintenance of that style code would also fall on their desk. That would be quite a large pill to swallow, especially given that most Zotero users are not in the law business.

    Looking at it from Zotero's perspective, robust legal support is relatively simple to implement at the technical level (and will be within reach when the schema is opened up for revision at Zotero 5.1 or so), but it also calls down the maintenance burdens noted above. Multilingual support is much harder at the technical level, and requires deep changes across the entire code base and the UI. As the Zotero UI will be completely reimplemented when the client jumps away from Firefox, it wouldn't make sense to put any work into multilingual until after the client is rebuilt to run on the Electron platform. After that, we'll see where things go.

    In a recent local development, we've begun work on a revamp of our faculty website, which will include extensive lists of (mostly Japanese) faculty publications supplemented with English metadata. Two assistants are using Juris-M 5.0 for the translation work, and we'll be pulling the results into the website build via the Zotero API, using some glue code to reconstitute the Juris-M metadata that we squish into the Extra field. The site rollout is scheduled for July, and we're all quite excited about it.

  • You wouldn't want to run Juris-M and Zotero in parallel […], but running them separately should be a problem.
    @fbennett : typo ?
  • Typo! Thanks for catching that one! Fixed.
Sign In or Register to comment.