Over the years I have come up with a system that work well for myself and in working with my students. When I collect research papers, I also download the publisher's citation in a bibtex format, including the abstract. I then group the citations into general categories, limiting myself to 5 or 6 subjects (e.g. drug design, nonbonded clusters, methodologies). The bibtex files are edited to have unique keys by taking the last name of the first author followed by the first letter of each following author's last name, and then followed by the publication's year (e.g. SmithMNOP2013). On the rare event that the same authors, with the same author sequence, publish additional papers that year, then I add an "a", "b", et cetera to the end of the year. The downloaded PDF files are subsequently renamed according to this key (e.g. SmithMNOP2013.pdf), allowing one to find the paper quickly according to the bibtex key.

When my student's follow this convention, it allows us to seamlessly exchange found information. We also have unique PDF directories set up on our networked home directories that are readable by members of the group. One could alternatively store these on a personal cloud service (e.g. Dropbox), which I believe maintains the wishes of publisher to limit the paper's distribution to the general public.

To easily manage the citations and PDFs, as well as to search the abstracts and titles for keywords, I use JabRef (http://jabref.sourceforge.net). This free java based program is operating system independent, and is continuously being updated by the opensource community. JabRef can also import citations that are in other formats (e.g. RIS). Given that the PDF files have the same name as the bibtex keys, it also can link to your PDF files once you specify where they are located. (Note that some care is needed since JabRef will complain about certain special characters, such as different types of dashes or Greek symbols.)

I have always had poor experiences regarding citation management and formatting when using popular Windows and Mac based writing program, including OpenOffice. I have switched to LaTeX because of this, and have found that switching the bibliographic formats as required by journals to be very easy now. There is a learning curve for LaTeX, but well worth it in the long run (i.e. a little time now save much time in the future). Most journals in my field now provide a LaTeX style file and submission options. I have also found that less errors are introduced by the publishers when typesetting the articles for publication.

Best regards,

Karl

Store your research source in pdf files in a new document library. Use software tools to look for themes in your sources. Try Qiqqa. Get it from http://www.qiqqa.com/70557

Ian I did not know this tool. I will evaluate it! After a first look, It seems more powerful than Zotero regarding the treatment of PDFs. The theme heuristics based in the research sources is very interesting. Thanks!

i personally use zotero and am very happy with the way it can import citation from chrome search through the chrome plugin... in addition it allows synchronization through the online library and it is very easy to share the files and citations between different computers..

pdf metadata retrieval allows easy citation by just adding a pdf file into zotero.

Thank you Ian for suggesting qiqqa.. from a look at the website there are so many additional and user friendly features that it is definitely worth a try..

I'm using Mendeley for PDF to store my references. But for bibliography i'm using EndNote. Since EndNote can be manually edited the references format, then it's easy to reformat the manuscript if the rejected manuscript to be sent to another journal.

just downloaded and tried qiqqa... naturally i had to compare it to zotero..

i added pdf stored copies of the following three articles to qiqqa and zotero..

1. Ooki A, Yamashita K, Kikuchi S, Sakuramoto S, Katada N, Hutawatari N, et al. Clinical significance of total gastrectomy for proximal gastric cancer. Anticancer research. 2008;28(5B):2875–83.

2. Varshney S, Manek P, Johnson CD. Six-fold suture: wound length ratio for abdominal closure. Annals of The Royal College of Surgeons of England. 1999;81(5):333.

3. Isreb S, Attwood SE. The fallacy of comparing surgeons with pilots in the search for safer surgical training. British Journal of Surgery. 2011;98(4):467.

while zotero was able to retrieve the citation from the pdf files in almost no time, qiqqa was not able to cite these articles stating the following reason...

"This document has no associated BibTeX information, such as title, authors, year and journal. Qiqqa InCite uses this BibTeX to help automatically build your bibliography."

this may be because i am new to the software and don't know how to properly utilize all the functions, and pdf metadata retrieval is only one function of a reference manager.. but i am impressed by so many other functions that are available in this software like a built-in pdf reader which is able to take snapshots from file and paste them on the brain-storming area... that has a lot of potential if properly utilized..

thanks again for recommending the software.. i am hopeful i'll find out soon how to use it properly..

regards, raza

Hi Raza I am making another tests. I have exported the sources Zotero to BibTex format with the documents associated. I started the import and its indexing the documents. I think that will be interesting in this tool its the expedition functionalities and brainstorming. At least is what I am interested for now.

I understand what you are saying, who uses Zotero likes is functionality of retrieving metadata for the PDF, I think that the equivalent in Qiqqa is "Qiqqa BibTeX sniffer".

Thanks for your feedback!

Hi Hahmad, what I like in Zotero is that we gives me the both options on a free tool... I never used Mendeley or Endnote, but the majority of my colleagues uses Endnote. Another advantage that I see in Zotero is its integration in Firefox and the easiness of saving a paper metadata and PDF directly from the publisher. Other advantage is the possibility to use with Dropbox and its availability for macintosh and windows.

But I am finding Qiqqa very interesting, because it can give me more possibilities, but doesn't work on mac :-(

Thanks for your feedback!

Thanks Pedro, Yes the BibTex sniffer does the job, although you have to manually select the citation for each article, whereas you can select multiple files in zotero which can retrieve metadata for all of them with one click.. But i agree the brainstorming function seems really powerful and i intend to use qiqqa for future projects..

Hi Hammad, Endnote is a good reference manager, but the strength of reference managers like Qiqqa and Zotero are that they are free..

Raza I have exported 3.5 Gb references and documents from Zotero and is still crunching the data, since last night, almost 24 hours :-)

My idea for now is to trying to use the both, and exporting the data on a regular basis , using Zotero autoexporter (at least I have a duplication of information, and if happens a problem in one application I can use another) . Another think that I am curious is about annotations export, I already used the annotations in Zotero with Zotfile.

I dont know if Qiqqa manages the annotations exported from Zotero, after the document OCR I will check it.

@Raza: Qiqqa has a BibTex Sniffer which will suggest to you the bibliographic info that you can then select. If you select multiple files or all the documents in a library, Sniffer will suggest for all in sequence. Auto retrieval of bibliographic info can be dangerous, for example, in declaring that the secretary who made the template or the document is the author.

Thank you very much Ian and Pedro.. This is quite helpful and I am really looking forward to getting a lot of work done with qiqqa...

i do understand the problems associated with auto retrieval of metadata, i always double check the citation with the actual article and only once i have found a totally different article cited for a pdf file and i had to correct it by putting in the PMID for that reference. Nevertheless, getting it done for many articles with a single click does tend to make one lazy.. :)

anyway i do intend to use qiqqa for doing it since i am very keen to try out the wonderful new features like brainstorming... thanks again for this very useful discussion and for suggesting such a useful tool..

regards, raza

@Julie: Qiqqa does not pull reference data from your file names. (I leave the file names unchanged, as it is less work, and easier to find the file again on the Web.) Qiqqa does use titles recorded in the BibTex info to create autotags and then use these autotags and your manual tags to identify themes. I had already arranged my files in different subdirectories, so Qiqqa used those as autotags. Use lowercase (singular) tags to avoid duplicating concepts. You can then explore your themes in Qiqqa Brainstorm!

Thanks Ian again for explaining this.. I was also kind of getting frustrated why qiqqa seemed to make a task out of sniffing citations for each file while it is so easy to pull metadata in zotero.. I did realize that it is in order to avoid a wrong citation easily possible based on how a file has been written. And i don't think any other reference manager uses the concept of autotags or provides so much space to actually work on your project and be creative. so i think i will become a convert soon.. :)

Hi! Normaly I work with Zotero auto renaming, after the reference is inserted in the database. @Ian I understand your idea, but in the other hand if the bibliography reference is defined, using for example IEEE Xplore and having its DOI available, I don't see much advantages in maintaining the file name, and personally I don't like to have 1234.pdf, and when I am working in Zotero after inserting a citation he automatically renames the file (ZotFIle PlugIn). The autotags and manual tags are very interesting to make some sort of "clusters analysis" to the tags, were I think that the predicting capabilities for the research areas is great.

@ Ian please correct me if I am wrong: Were I think that Zotero is better, is when we are in the web accessing articles databases in a browser (since Zotero could be installed as an extension in Firefox), e.g. scholar, IEEE and insert directly the citation or when you go to LOC or amazon and create a book citation directly. In the other areas I think Qiqqa wins almost in every aspects.

But Zotero works on Mac, windows and Linux and for me is good the options of windows and Mac.

@Jane interesting, Papers was one of my first options, I am not sure, but I think at the time windows hasn't available and I didn't adopt it. I like Zotero combined with Dropbox. I have my documents on the cloud and a lot of spaces to do so.

@Julie in my school I have a similar case, the tool more used its endnote :-) I rename the files also as you commented, basically use a Zotero extension Zotfile for that, it recreates the filename based in the author name, year and article name, but can be changed. I agree Qiqqa its more advanced for research analysis.

Sometimes we have to complete manually the article information, normally I search in scholar but I try to complete using catalogs or publishers, here Zotero inserts a symbol in the URL that facilitates the insertion of the reference and document (if available). This functionality its pretty cool! Also available in loc.gov, amazon.com, scholar ....

We have to thank to @Ian for its suggestion :-)

Over the years I have come up with a system that work well for myself and in working with my students. When I collect research papers, I also download the publisher's citation in a bibtex format, including the abstract. I then group the citations into general categories, limiting myself to 5 or 6 subjects (e.g. drug design, nonbonded clusters, methodologies). The bibtex files are edited to have unique keys by taking the last name of the first author followed by the first letter of each following author's last name, and then followed by the publication's year (e.g. SmithMNOP2013). On the rare event that the same authors, with the same author sequence, publish additional papers that year, then I add an "a", "b", et cetera to the end of the year. The downloaded PDF files are subsequently renamed according to this key (e.g. SmithMNOP2013.pdf), allowing one to find the paper quickly according to the bibtex key.

When my student's follow this convention, it allows us to seamlessly exchange found information. We also have unique PDF directories set up on our networked home directories that are readable by members of the group. One could alternatively store these on a personal cloud service (e.g. Dropbox), which I believe maintains the wishes of publisher to limit the paper's distribution to the general public.

To easily manage the citations and PDFs, as well as to search the abstracts and titles for keywords, I use JabRef (http://jabref.sourceforge.net). This free java based program is operating system independent, and is continuously being updated by the opensource community. JabRef can also import citations that are in other formats (e.g. RIS). Given that the PDF files have the same name as the bibtex keys, it also can link to your PDF files once you specify where they are located. (Note that some care is needed since JabRef will complain about certain special characters, such as different types of dashes or Greek symbols.)

I have always had poor experiences regarding citation management and formatting when using popular Windows and Mac based writing program, including OpenOffice. I have switched to LaTeX because of this, and have found that switching the bibliographic formats as required by journals to be very easy now. There is a learning curve for LaTeX, but well worth it in the long run (i.e. a little time now save much time in the future). Most journals in my field now provide a LaTeX style file and submission options. I have also found that less errors are introduced by the publishers when typesetting the articles for publication.

Best regards,

Karl

Hi Karl I liked your suggestion, and if I continue to work in research after finishing PhD - I hope so :-) I will consider It. But for people who don't have good skills in Informatics probably this kind of applications its more easy. At the end

I liked your systematic approach for handling documents and references, I am sure that I will try adopt it with my students, its a clean approach and simple, and can be formulated in easy steps as you proved.

I got curious and for future references were you write the articles? Specific application to handling LaTex? Or a regular text editor, like emacs?

It remembered the early stages when I used Tasword for text editing in an Amstrad 1640, without graphical interfaces - sometimes i miss those times :-)

Thanks!

Pedro

@Pedro: Local file names should be whatever works for you. I'm just deliberately lazy.

@Pedro: Qiqqa has a built-in browser, or you can specify Firefox or Chrome and pick up the details from Scholar etc. using Qiqqa's Sniffer straight into BibTex format.

No, Qiqqa is not available for the Mac, but it is available for the Android OS also.

@Karl: The troubles with limiting yourself to 5 or 6 subject areas (libraries in Qiqqa language) is that you cannot easily store a paper that deals with 2 or 3 subject areas. Qiqqa lets you tag each file with more than one tag, and 1. Search over that tag or tags. 2. Search in a library (subject area) for a word or phrase anywhere in its documents-- not just the biblio info. 3. Search over all your libraries (subject areas) for a word or phrase anywhere in their documents. Search results are shown in context. Finally you can Brainstorm any library (subject area).

@Karl: This is a very unique and very useful method of file storage. Thank you for sharing it. I normally name the files by the article title but it gets too long but this seems a very nice and useful system that everyone can easily follow...

Ok! Qiqqa finished the stress test :-) Considering that all the information that I find relevant, including from practitioners views and books I insert them in Zotero, basically I use it for research and personal library - Its my approach to treat the information and I have more than 1000 entries, much of them annotated.

The process of exporting the database and documents from a windows machine from Zotero was unworkable. The export to BibTeX format with the docs attached always crashed, I managed to do this exporting from my Mac.

The process of importing the information on a windows machine went during almost 48 hours, crunching the data (processing PDFs). I have done this in PC not very powerful, and I had several crashes in the OCR.

Some considerations to Qiqqa:

(1) Search functionally in Qiqqa its great, considers authors, relevance, et cetera. Its a must have;

(2) Maintained the annotations and text highlights, this for me was great;

(3) Running expedition in one reference, makes the suggestion based in the recommend themes and articles relevant - also useful;

(4) Themes generated some Ok others not, probably needs some tuning in "Manage AutoTags" or making an export more delimited to the research field were I am working;

I had the references in Zotero organized in folders, and I imported everything to a library, probably needed to divide it using libraries more subject related or export more delimited. Now I feel a little bit lost.

I think Themes treatment its fundamental in Qiqqa, for the suggestions and relations between papers become more smooth. Suggestions for this area will be great!!!

Thanks for your opinions, I think that we had nice information handovers here, that I am still digesting.

Best regards,

Pedro

wow Pedro, thats some extensive work... very impressive... i agree there are certain strengths to the qiqqa platform in addition to just storing references and citation that others don't have, especially working with/analyzing the references and progressing with the research...

one of my areas of interests is qualitative research and i am thinking qiqqa can be easily used to go through pdf transcripts and the grids/analysis can then be performed in the brainstorming or the editing mode making the job rather quite easy..

it has been a very useful discussion and i look forward to more useful tips and experiences from qiqqa..

regards, raza

@Pedro, I agree that LaTeX may not be the best way to go for folks who have not been exposed to programming language. However, I believe most people can learn LaTeX given some effort and time.

There are several programs that are usable by different OS which can be used to write/manipulate LaTeX documents. Being a Linux user, I prefer Kile. Kile has a nice GUI with many customizable options and LaTeX commands built in. It will also compile LaTeX and bibtex documents and allow you to view the resulting PDF. (Emacs also works fine.)

If you are writing your PhD dissertation, I would encourage you to do so using LaTeX. (Obviously, it is up to you, your adviser, and your institute with how you generate your dissertation; plus time issues always become important at the end of your PhD education.) You can create a single and simple master LaTeX document that references separate LaTeX documents. These separate documents would be individual chapters from your thesis. In doing so, you will be able to manage your citations, tables, figures easily. It will also automatically create a Table of Contents, a list of tables, a list of figures. If you have math formulas, then LaTeX will format them very nicely.

The key to do this, is to compile the document every so often (5-10 minutes or after significant editing). This will allow you to trouble shoot easier if the document does not compile, which simply be due to a typo (e.g. forgetting a \ ).

@Karl I think that I understand the approach to write a big document... I will try, probably with a test... Another aspect that interesting its that the writing is broad used and could be integrated in several environments.

I didn't realize that its a markup language for documents (like the HTML for the web) with option to prepare the documents, that could be interpreted in several environments. The layout could be generated in other places, like we do with CSS for HTML.

Definitely I got very curious about it and the possibilities of reusing the same contents with different outputs...

Thanks Karl!

@Pedro: OCR of files containing only a diagram will not be successful. The same applies to strange fonts (comic!).

(4) Themes are suggestive, to help you brainstorm, make new connections and organise your thoughts.

You do need to use the AutoTag manager to blacklist "stupid" acronyms like ICT or common English phrases like "at this point in time"!

It also makes sense to export only one of your projects to one Qiqqa library. You can still recover gracefully now by using Qiqqa to move a project's pdfs to a new library.

You can change the number of themes to retrigger a new classification of themes, which may be marginally better...

But don't do it accidently if you have found a beautiful pastel rainbow clustering!

Further refinement of clustering will occur once you tag each PDF file with a couple of your very own interest tags.

@Raza: You can put all your transcripts into one Qiqqa library, and look for emerging threads!

Start off with autotags, then go on to tag each file with a number of your own special tags e.g. one per grid column and/or one per grid row. I have tested Qiqqa today, and it works for my test case of 26 tags for one document.

Thanks Ian, sounds like a very workable idea and i will start doing that on qiqqa.. So far it has been working really well for me but i have not started working on the brainstorming yet.

would look forward to further tips and support in future..

regards, raza

Ian I get you point, use it as data mining to generate patterns operationalized in Qiqqa as themes. The black list is fundamental I had some generated themes that are a nonsense. Is there some black list of words prebuilt? Or the solution is to go to nonsense themes and add it to the blacklist?

I don't know if I understand this "You can change the number of themes to retrigger a new classification of themes, which may be marginally better..." are you referring to the number of themes (by default) to be generated? What do you mean for marginally better?

The themes can be improved adding manual tags to the papers? Doing this doesn't mean to have also impact and improve also nonsense generated themes?

Thanks Ian!

Following the suggestion of Julie openned added a question to ideas sharing for Qiqqa at https://www.researchgate.net/post/How_manage_research_with_Qiqqa.

Ian I hope you agree? Because you have been the mentor and we don't want to loose your feedback :-)

Yes Julie and Pedro, it certainly makes sense... i think i like the idea of additional manual tagging... but i haven't worked out how to 'blacklist' or 'control' unwanted autotags... i guess we all will figure out what suits our indivual 'qiqqa' style.. :)

i think the best thing about qiqqa is that it gives your research work your individual touch based on your preferences, and i'm very happy with this quality...

@ Ian: i totally agree with Pedro that we would like your guidance quite frequently in this matter...

regards, raza

@Pedro: In Qiqqa, there must be a list of built-in stopphrases and words (such as "This research" and "This" in English"). What Qiqqa out-of-the-box does not know is what the garbage "common" phrases are in your specific collection of documents. So, simply blacklist any that you see listed in a theme, and re-explore.

The themes should also improve noticeably if you add your own tags, because Qiqqa is then being taught the way you think about the domain.

For obvious reasons (if you sit and think about it!) Qiqqa tries hard to put an equal number of documents into each theme, so by changing the number of documents in a particular document library or (more artificially) by changing the number of default themes to be generated, a re-exploration will give you a new, different clustering, which also may be closer to the way you think (i.e., "marginally better").

By re-exploring with only autotags on, you see the themes that the documents are telling you about; by re-exploring with only your tags on, you see the themes that you are looking for from your viewpoint.

I usually re-explore with both autotags and my tags on simultaneously, to get the best of both worlds.

Because you are using your own controlled vocabulary in tagging, you will not introduce any nuisance terms (otherwise you are mad!).

@Julie: You will not necessarily find your tags listed in your Qiqqa

themes. For example, your themes may be "AAA", "BBB"... "ZZZ".

What if those words occur in none of your documents? I am not sure if your OCR of Greek documents is working. You can test by searching for a Greek phrase.Nor am I sure if you are using English tags on successfully OCR'd Greek documents.

You may find that the manual tag "Greek" will be useful for you.

Qiqqa is trying (very hard) to show you themes (and the probable documents that you should first look at when writing a review of that theme).

I regard a tag as being something significant that you believe is in your documents and which you want to coax out (tease out) of the

document.

I regard an autotag as being something significant in your document library (from a statistical viewpoint). This may be the acronym

for the institute that wrote each chapter of the book which you

have stored as one chapter per file. It may or may not be

a useful trigger for looking for patterns in your documents.

I regard a theme as being a section in your research paper or a

chapter in your thesis-book. Qiqqa will suggest to you possible

subdivisions (themes) which you can use to structure your writing.

@Raza: How to blacklist: In Qiqqa's Expedition menu there is an icon in the form of a blue ball called Manage Autotags.

On each line in the right column, add an autotag that has appeared

in a theme, but which you never want to see again.

Hello, I am the founder of Docear which is a free and open source literature management software for Windows, Mac, and Linux. Docear has a unique concept that helps you organizing, creating and discovering academic literature with a strong focus on the management of PDFs and annotations.

If that sounds interesting to you, have a look at http://www.docear.org/2013/10/17/docear-1-0-stable-a-new-video-new-manual-new-homepage-new-details-page/

And you might also be interested in this article which is discussing some aspects of how to (not) chose a reference manager http://www.docear.org/2013/10/14/what-makes-a-really-really-bad-reference-manager/

Joeran thanks for the suggestion, I will have a look and test it.

We have been articulating the discussion mainly around Zotero and Qiqqa, with more emphasis in the great features of Qiqqa. With the focus on Qiqqa content analysis features, that permits for e.g. generating themes automatically to organize the papers by research areas (themes), which is something that is not available in Zotero.

Nonetheless, Zotero has some features for content annotation (PDFs) and facilitates the notes extraction using additional plug-ins!

Without a test, it seems that it could be in the middle, between Zotero and Qiqqa in the area of references management. And it is free :-)

Pedro, unfortunately there is no reference manager that has it all. You always need to compromise in some way :-( I would recommend Docear for those users who want to heavily annotate their PDFs (i.e. highlight text, create comments, or bookmarks) and keep an overview of the PDFs and annotations. I hope you will like it :-)

if you are interested in a comparison of Mendeley and Zotero (and Docear), our new in-depth review might be interesting http://www.docear.org/2014/01/15/comprehensive-comparison-of-reference-managers-mendeley-vs-zotero-vs-docear/

It is great to have different reference managers that can be used freely and selected according their characteristics and individual needs of the users.

Thanks for the links and suggestions!

Very useful Topic 

Thanks for the links and suggestions!

Thank you Joeran Beel. 

For me Citavi works well (https://citavi.com/de). Its free until 99 publications.