Mendeley

I started to play around with Mendeley, which offers sharing of PDFs and references, as well as annotation features, very much like Zotero in fact. I'm not sure why, but I decided to stick with Zotero, especially because I've been using it for the past few years now and I very much like the way in which it developed. I also have the impression that Mendeley is not open source.

However, the web interface of the Mendeley library and the annotation features, as well as the profile features look really great. I hope Zotero moves in that direction.
«13
  • My impression is that Mendeley will take over. They are a bunch of super fast guys. They have almost a monthly new version.
    Zotero should definitely try to associate with them. Zotero is still ahead on single-click-in-your-library, but their product is or will soon be superior. I regret to say that because I am definitely for the opensource option.
  • CB
    edited March 18, 2009
    DNAcombo: could you expand? What features does Mendeley add or do better (I haven't tried it)?

    [edit: the word I've just got from a colleague who tried an earlier version is that its metadata extraction was poor, and she ended up manually entering most item details. ]
  • edited March 18, 2009
    @DNAcombo: that's an entirely superficial assessment.

    If you're an historian, or a legal scholar, I'd venture to say that Zotero is far ahead of Mendeley. And the open source nature of Zotero, while not fully exploited by the Zotero team (which I think holds them back) is no trivial thing. You need to ask yourself, what happens to your data and your documents if Mendeley''s business model fails?

    OTOH, Mendeley may have some current advantages if you're in the sciences and big on the collabroation features that Zotero has long promised, but not (yet) delivered.

    Here's the thing, though: both Zotero and Mendeley have the same basic vision. They just have slightly different plans to get there.

    For that reason, I actually think there's a lot of room for useful collaboration, and have told both developers the same:

    1. fully open data sharing and exchange; RDFa and bibo support is a really easy way to go a long way with this, but I've yet to see it in either project)

    2. CSL integration; both support it, but there's room for improvement

    3. implementing a standard API in word-processors so that we can have real interop here

    Rather than getting into a pissing contest about which project is better, then, I'd like users to press the developers of each project to actually deliver on the openness that is the foundation of both of their visions. There's no reason both projects can't flourish.
  • CB
    edited March 18, 2009
    On:

    the openness that is the foundation of both of their visions
    Isn't Mendeley a for-profit closed source app (not that I'm particularly doctrinaire about such things, just interested)?

    [edit: though this seems to suggest they are open to collaboration]
  • Yes, it is (though they've suggested there's room for open sourcing at least part of their stuff).

    I suppose I'm taking a particular interpretation of what they're up to, suggesting that it can't succeed unless it's based on open data and open standards, that's it's certainly not in their users' interest if it's not, and so framing expectations ;-)

    More practically, I want to see them work together where they can (on the issues I note).
  • Well, kudos for your various attempts at encouraging openness and decoupling (it's quite a battle. My institution has recently enforced a new hosted email system that requires students and staff to sign up to an obnoxious fake 'agreement' with Microsoft. Such moves seem to be gathering force if anything).
  • Perhaps it would be more helpful to clarify which parts you want open-sourced? The software code or the data? And the reason for doing so. I'm failing to see the logic. I think Mendeley's intention is to open up the data via APIs (with consideration to privacy of course).

    I don't see why they need to open up the code to the software though, same would go for Zotero. Do you demand the same of Google Scholar? Apple? Microsoft? It would only make sense if they run out of funding, otherwise investors would hardly be willing to continue financing the project. This isn't like MySQL or Ubuntu that started from a single altruistic individual or group of developers. There are a few vocal supporters of open-source, but the majority of users will not care as long as data and privacy are not abused and open-source does not solve that problem.
  • DocGene: I think you've scanned the thread a bit fast. I don't see any demands here that Mendeley open their source. The strongest prefs expressed here are about openness re sharing data, and Bruce suggested that the Mendeley developers are showing promise in this respect.
  • I don't see why they need to open up the code to the software though, same would go for Zotero.
    As this is your first post, I question the motivation a bit. However, I'll bite.

    Open source guarantees the program will be available, at least as-is, in the future. Mendeley is currently available at no cost after you register. But they could change this, and decide to charge for their program. Or they could go out of business. You would no longer be able to download their program.

    Open source allows other to submit bugfixes and to add features. This is already happening with Zotero. Mendeley has said part of their business plan is to charge money for "premium features." Because the program is proprietary, there may be no way for non-paying users to gain access to those features.

    Open source allows usage on more platforms. Mendeley is better than most proprietary apps in this regard: they offer their program on x86-(32,64) Linux, for example. But I run Zotero on ppc-Linux and FreeBSD and other platforms. I'd be able to run it on future platforms too.
    Do you demand the same of Google Scholar?
    I don't give Google Scholar data or rely on it as my only source of data.
    Apple? Microsoft?
    Ditto, but also note these companies are MUCH larger than Mendeley, so the threat of them collapsing is much lower.
    It would only make sense if they run out of funding, otherwise investors would hardly be willing to continue financing the project.
    Given the current financial outlook & the huge risks in any for-profit software startup, you shouldn't dismiss this possibility.
    This isn't like MySQL or Ubuntu that started from a single altruistic individual or group of developers.
    Neither of these were altruistic ventures. MySQL AB and Canonical are for-profit companies.
    There are a few vocal supporters of open-source
    And I'm obviously one, but you'll find others here too.
    but the majority of users will not care as long as data and privacy are not abused and open-source does not solve that problem.
    open source does make the data as open as it could possibly be, though.
  • noksagt: the risk argument is often raised against closed source. But in practise the risk is in proportion to openness regarding data. I don't find it unacceptably risky to use closed-source gmail, because it supports IMAP and POP, so I can extricate myself instantly if I don't like their direction. If gmail didn't offer at least one of these, I wouldn't touch them with a barge pole.
  • CB: Note that the risk argument was not the only one I raised. It is still valid for products that use "open APIs" or allow "open file formats" to be exported from them, though.

    What happens if you forget to or can't export before the ship has sunk? Google is "too big to fail," but POP & IMAP would do you no good if you hadn't already downloaded all your mail through one of them if their servers went down.

    Again, Mendeley is better than most. They also use sqlite for their backend, so their core data is more accessible than, say EndNote's. But all you will have if something unexpected happens to Mendeley is your data: you will have to start from zero to be able to use that data. You'll at least have to sit down & figure out the schemas for the database & figure out how to get it into some other program that you'll be able to use. Again, this isn't the worst thing in the world. But it is certainly more costly than what happens if GMU stopped developing Zotero.

    There's value in open data. Mendeley isn't a bad product. It is much more friendly to end-users/community than most other proprietary software. (If anything, this makes me a bit more paranoid--I can name dozens of proprietary apps that have been available on Linux that have failed. But I give them credit where it is due.)

    I'm just saying there is also plenty of additional value in open source as well.
  • CB
    edited March 19, 2009

    What happens if you forget to or can't export before the ship has sunk? Google is "too big to fail," but POP & IMAP would do you no good if you hadn't already downloaded all your mail through one of them if their servers went down.
    True, but that's a function of using a hosted solution rather than of software licensing. I'm not going to run a mail server.

    But all you will have if something unexpected happens to Mendeley is your data: you will have to start from zero to be able to use that data. You'll at least have to sit down & figure out the schemas for the database & figure out how to get it into some other program that you'll be able to use. Again, this isn't the worst thing in the world. But it is certainly more costly than what happens if GMU stopped developing Zotero.
    Actually in either case I'm stuffed, or at the very least am hostage to a third-party doing something about it.

    As a user, all I really want is for my data to be permanently available to me. I can think of only two types of data where this has been consistently the case over the years: email and text files (ie. I have usable and permanent archives of both of these). With neither of these do I have a pressing interest in what app accesses them (or what the license for that app is). Everything else has suffered some form of attrition.

    [Edit: this is just to say that the fact that Mendeley isn't open source wouldn't be a big factor in evaluating it, for me. There are plenty of arguments for open source in all kinds of arenas.]
  • @noksagt - I saw a link to this post via Twitter and since I am a user of EndNote, Papers, and recently Mendeley I thought I'd chime in. Not trolling. I'm just a believer that not everything needs to be open (in this case code) and that science-related organizations have the right to be for-profit and this can be a good thing.

    Though I do advocate for open data and other than EndNote, most companies in this sphere seem to agree. Open data will drive innovation and technology, more so than open technology...in theory.

    Thanks for the response.
  • @CB - You may have been correct about scanning the thread too quickly. Apologies.
  • edited March 19, 2009
    @DocGene: fair enough, but let's turn the question around: why do you, as you say, believe that "not everything needs to be open (in this case code) and that science-related organizations have the right to be for-profit and this can be a good thing"? You say you believe this, but why? With the collective knowledge and experience that the scholarly community has, why shouldn't we create and sustain these tools ourselves, with complete control over their evolution and long-term viability?

    The "free" in "free software" is about much more than cost.
  • I don't think an open source discussion here will lead anywhere - some people don't care, other's feel passionate about it - I don't think noksagt is quite fair - I think the chance that mendeley will go down from one day to the other is very small, so you could almost certainly extract your data. I personally think there are good reasons for open source especially in the scholarly community - and the way some of the os scholarly projects are outperforming their competitors (R, LaTeX) shows that - but, more to the point:

    about Mendeley specifically, having looked at their features:
    I think it's clearly inferior as a citation tool, which is quite an important function of Zotero for many here (I don't think, for example, you can change the citation style in an existing document. Can you even auto-generate a bibliography? )
    If you say it's also inferior as a "one click" in your library - then it's really just the social networking and metadata extraction (does that work a lot better?) where it's better.

    I just don't get the sense that it's being developed with the same knowledge of the scholarly process that's one of the great advantages of Zotero.
  • @ adamsmith: I totally agree. Similarly to what Bruce mentioned above, Zotero is far more interesting, at least in the humanities and the arts, also because there are far more item types available in Zotero. Second, Mendeley does not extract meta-data from library catalogues and other websites. As a reference manager or research tool, the Mendeley desktop is much more limited. To give a few other examples:

    Advanced power search in M: No; Saved/smart searches in M: No; Cross-referencing or relating items/notes in M: No; Create sub-collections in M: No; Add/create citation styles in M: No; Generate reports in M: No; Snapshots in M: No; Etc.

    Yet, as I wrote in my initial post, its online interface seems very promising. I'm sure the Zotero team will eventually incorporate similar features, and perhaps move beyond them. They have always surprised us :) I have a long history with reference managers and Zotero is the first (and still the only tool) that gives me intuitively the feeling that it is developed by scholars, for scholars.
  • Yes, good point: people can consider my question rhetorical.

    But the issues aren't just about philosophy; they are practical too. For example, I've not looked at Mendeley in detail, but will trust raf's account. That's not good.

    Zotero, OTOH, has pretty good, in some places great, word-processor integration. But that functionality is specific only to Zotero (and currently, only to specific Zotero users).

    I recently wrote a post about the extreme lack of interoperability of citation processing tools WRT to documents.

    So one obvious question is how, practically, do we fix this mess?

    Well, one obvious solution is what Thomas Zander (a lead KOffice developer) appears to be proposing in the comments to that post: collaboration on open source code for a generic word-processor plug-in and/or API.

    This may be a longer term project, but as a practical point, I don't see how these problems get solved without the help of open source development. I don't see how it's practical for projects to develop and maintain their own unique code for this.
  • Hi all,

    Victor from Mendeley here - I thought I'd chime in! Thanks for all your comments, we do appreciate them.

    As Bruce has pointed out, there are similarities in Zotero's and our features, and both of us are already using Citation Style Language (CSL) to format references. Zotero is a great tool, especially for grabbing citations off the web. As far as we're concerned, we'd be happy to collaborate, e.g. (as Bruce suggested) to support a common standard with citation styles or to make our respective word processing plugins interoperable.

    For our next major release (due in about six weeks), we're also planning to develop a two-way integration between Zotero and Mendeley - if you have both installed, you should be able to sync your references either way. It would be nice to get the thoughts of Zotero’s team on this! Hopefully, the CiteULike integration should also be up and running by then, so you should even be able to sync your references between all three services.

    Finally, allow me to respond to a few comments concerning Mendeley:

    "(I don't think, for example, you can change the citation style in an existing document. Can you even auto-generate a bibliography? )"

    You can indeed change the citation style in an existing document - since we use CSL, we support the same styles as Zotero does. Multiple Mendeley users can also collaborate on the same document and auto-generate bibliographies when using Mendeley’s “Shared Document Groups”.

    "Advanced power search in M: No"

    Our interface doesn't show it (yet), but the Mendeley Desktop search field does support Google-style advanced queries like "title:searchterm", "author:searchterm", or placing the searchterm in quotation marks to search for a specific phrase. Obviously, we need to make this clearer.

    "Saved/smart searches in M: No"

    Indeed, you can't save searches yet. As for "smart" searches, Mendeley Desktop automatically generates author, journal, keyword and tag filters which can be used for searching.

    "Generate reports in M: No"

    If you use Mendeley Web to sync your documents across computers or manage your reference online, Mendeley Web does generate reports/statistics – maybe not the particular ones you're looking for, and not yet as exhaustive as they could be.

    Hope that was informative!

    Best wishes from London,
    Victor
  • Hi Victor (and everyone else in the thread) I'm not a Zotero developer, Im just an academic who is a fan of open source/free software who currently uses both Zotero and Mendeley. I must say that its great to hear that you guys over at Mendeley are willing and eager to interoperate with Zotero and play nice with the free and open crowd (were used to getting sand kicked in our faces by corporate bullies like endnote).

    Given all the exciting things that are still to come with both of these citation managers / academic communitites, I would love to see an effort from both Zotero and Mendeley to work on compatibility and interoperability for not just the citation databases and office plugins but also for the online communities and all the wonderful things that will come from that. That would be truly exciting and terrific.

    So Victor and Zotero Devs, what are your thoughts on eventual web service interoperability/compatibility (e.g., shared networks, files, etc..)???

    Thanks and take care,

    Pat
  • Well, the biggest advantage of Z over M is the way it can be embeded into my webbrowser - no additional programmes etc. And most important for me is also the way it handles proxies (using Ezproxy to access university network from outside).


    ---- spoiler ----
    However, the biggest drawback is that Z tends to lose already saved proxies after each update of the Firefox add-on...
  • Earlier versions of Zotero weren't saving proxy information correctly, and it was necessary to purge that information. Future updates should not lose saved proxy information. Apologies for the inconvenience.
  • @ noksagt
    "What happens if you forget to or can't export before the ship has sunk? ... They also use sqlite for their backend, so their core data is more accessible than, say EndNote's. But all you will have if something unexpected happens to Mendeley is your data: you will have to start from zero to be able to use that data."

    I didn't try Mendeley (yet), but from what I've seen on their site I understood that I can have both the software and my data locally. And should the company fail (and I really wish them success), I'll still be able to continue to use the software and the data. Or export them to Zotero. Or did I miss something about Mendeley?
  • I just installed Mendeley today after reading about it on Lifehacker. I'll keep trying with it but first impressions were not good. I found it a pain to install; there appears to be no help function; the help refered me to a pdf not included in the install package; the plug in for open office crashed a few times; it doesn't do any of the useful data scraping I use zotero for.
  • jonbotyson: there's no doubt Mendeley is a couple of years behind zotero in maturity. But it has a different approach, and is very promising in its own way.

    On your specific points:
    The pdf help file should be there in your default library. Mendeley's approach to getting metadata is different from Zotero, but actually works quite well once you get a handle on the various methods (web-based bookmarklet + pdf metadata extractor + DOI or PMID lookup). I don't use any of the wp plugins in either zotero or mendeley so have nothing to say about them.

    Mendeley does have one big advantage over Zotero as things stand, and that is that the web app has full functions (ie. isn't read-only). This has driven me to use it for a current project, where I need to add to my library from computers I can't install anything (firefox extensions included) on.
  • I've tested M, mainly for just for fun and to know...

    negatives:
    - it failed to load the Library Z exported in UTF-8 or any other encoding
    - it stuck on some syncing or extracting, I had to kill it a couple of times
    - it allows 'related' but it's not so straight-forward as in Z

    positive:
    There is the sharing option, but Z has the RSS (it just does not work for me as expected, GReader failed to load...) and that's fine with me. Anyway in our group we do not want to share anything anybody sees and saves, I'm not interested in that. But when somebody saved something interesting I'd like to know about it and then decide on saving or not.
  • @ben58
    You haven't missed anything. Your Mendeley data syncs with citeulike and Zotero, is backed up online and locally, and is available for export in a variety of formats. You have plenty of options if something were to happen to Mendeley as a company.

    @jimbotyson
    There were some issues with the linux and mac versions, have you tried the latest release? The developers are very available via twitter (@mendeley_com and @mendeleysupport)

    @zroutik
    Mendeley now has RSS feeds for subcollections which you can embed or syndicate as a public collection and there is also a "shared collection" feature which allows you to share PDFs with their annotations and metadata with a selected group. This is great for journal clubs and the like.

    I'm available on twitter as @mrgunn as well. I serve as an academic community liaison for Mendeley.
  • edited September 27, 2009
    @williamgunn:
    Your Mendeley data syncs with citeulike and Zotero, is backed up online and locally, and is available for export in a variety of formats. You have plenty of options if something were to happen to Mendeley as a company.
    Two issues related to this:

    1. Last I checked, Mendeley had no open API, so just how much of one's data can one easily extract?

    2. Documents you create in Mendeley only work with Mendeley (a failure based in part on the fact that they've used Zotero's code for this stuff).

    The two of these together means the costs associated with moving to other solutions should "something happen to Mendeley" are higher than you suggest.

    Note: not saying Zotero doesn't have some of the same issues, but I really want users to keep these issues in mind.

    Also, not sure about your claim that Mendeley data syncs with ... Zotero; how?
  • @bdarcus

    The open API is coming very soon, I hear. The Zotero sync is currently only available in the dev preview, but it's open to the public and stable on my Windows machine. It's also only one-way sync from Zotero, but that may change soon.

    The nest point you raise about there not being a standard for word processor integration is a good one, but not one Mendeley can solve on their own. It's going to be an important issue in the future, so I hope they can work together on this.
  • hmmm I've tried the Medeley/Zotero sync and so far all I get is a bunch of empty collections... Obviously I'll ask over at the Mendeley forums...
Sign In or Register to comment.