We often need to also read the original papers cited by the article i'm currently reading - especially when reading reviews or when reviewing an article pre-pubication. it would be SUCH great help if Mendeley would let me select certain papers and source all the citations used in those papers for me!

When automatically renaming PDFs, sometimes file names result too long (with potential compatibility issues). In the PDF renaming menu, you can add:
- a Field to determine the max characters of file names
- Options to replace blank spaces and special characters: so f.ex. that ' ' turns into '_', and 'á' turns into 'a'

In automated naming conventions I am using author, year, title. Using this convention, titles get excessively long. It would be good to either automatically detect the file system or let the user determine the maximum length in characters for a file name. I have run into issues of putting files into subdirectories or the trash on NTFS as a result of names being too long.

Mendeley seems very poor at extracting metadata from articles downloaded from Jstor, presumably becasue of the additional Jstor page tagged on the front.

Perhaps Jstor articles could be recognised as such, and parsed differently?

It would be great if we can have an option for working offline. This will give us peace of mind that we don't accidentally share the stuff that we are not ready to share (by clicking on the wrong links or buttons).

To improve the social aspect and usefullness of Mendeley, it would be cool if there was a conference calendar, where you could show people what you were up to, and where. This could be expanded to show on a global map, where conferences are happening, make suggestions to which conferences may be interesting to check out etc.

Eventually it could be a good idea for conferences to make their own Mendeley conference groups holding invited papers, presentations etc.

Displaying a tag cloud in the desktop application would be nice.

Similar to MS word's split view feature, so that a bit of the screen can be used to look at the references while the rest can be used for reading the paper, so that can see what the references are while reading without having to scroll all the way down and then back to where you were reading

Although the web importer is awesome, it is cumbersome to use it when exporting several references from PubMed. PubMed can export several references to a .txt file in 'MEDLINE' format, which is not exactly a RIS format that Mendeley can handle. Importing this kind of text files would be a great tool for Mendeley desktop. Thanks.

I have many Ph.D. theses in my document library. Some of them are over 50MB. I'd like to have them stored with the rest of my document library in Mendeley.

This would pull up a list of citations in the article currently being read on the side (so the reader wouldn't have to scroll down mid-read to find them) and give the reader the option to download these directly.

The automatic metadata extraction form NASA ADS needs improvement - essentially every paper I import I have to go in and manually add a DOI to get the important data included despite that fact that it's very easy to get ADS entries including the DOI in an easy-to-parse XML format (just add ?data_type=XML to any article). Furthermore, keywords are never imported, even when they are clearly present.

It would be great to be able to have the ISI impact factor of journals displayd

Hi, why not easily adding synchronization with BibSonomy (http://bibsonomy.org/) like you did with CiteULike?

When the title of the paper has some special characters (such as ":" or "$") Mendeley cannot save it proprely, it would be nice to fix that problem