Take the 2-minute tour ×
Academia Stack Exchange is a question and answer site for academics and those enrolled in higher education. It's 100% free, no registration required.

When skimming through a long mathematical paper (in pdf format), I often find a mathematical symbol (often in greek and with subscripts/superscripts) that was defined previously somewhere else within the paper. Of course, it's not always easy to find where the parameter was defined, so I resort to using the FIND tool. However, simply copy and pasting the symbol to the FIND tool doesn't seem to work well, since the symbol cannot be copied exactly.

Is there a better way to perform an automated search a PDF document for a particular mathematical symbol if I don't have the original latex file? Especially ones with subscripts and superscripts?

share|improve this question
1  
What PDF viewer are you using? Using Evince, searching for symbols seems to work without any problems. –  Senex May 18 at 14:17
    
@Senex: I'm currently using Adobe Reader. I've tried Evince, but it doesn't seem to work well to search symbols with subscripts and superscripts. Anytime I try to copy such a symbol, it doesn't paste exactly in the FIND tool. –  Paul May 18 at 15:33
1  
OK, I see the problem with superscripts AND subscripts (with only one -- either a superscript or a subscript -- Evince seems to work by copying and pasting the symbol and deleting the space between the symbol and the sub/superscript that gets added during copying and pasting). –  Senex May 18 at 15:39
    
I search for '?', papers seldom have many questions. The precision is bad, but the recall is good, in my experience. –  Trylks May 19 at 15:26
    
Are the equations not numbered? –  Jonathan Landrum May 19 at 18:07

2 Answers 2

This can be a tricky one, and you can sometimes have similar issues with accented and greek letters. If you cannot find words that have been hyphenated to break a line, in my experience this is reader-dependent and Adobe Reader has never given me problems.

Some PDFs just were not generated with the proper character mappings, so they will never be searchable with the affected characters. With regards to the format of superscripts and subscripts, the PDF will be searchable with them in the same format that they appear when you copy-paste out of the document. With other special characters, if you cannot copy-paste them, you cannot search for them.

If you urge everyone using LaTeX to do \usepackage{cmap}\usepackage[T1]{fontenc}, those packages should solve most problems. Notably, cmap makes math characters (such as the integral sign) searchable.

Unfortunately, as far as I know there is no way to "repair" a PDF that is not properly searchable, unless you re-create a fixed version from the source.

share|improve this answer
    
If it were searchable, then simply copy-pasting from the PDF file should work, right? –  Paul May 18 at 16:42
4  
@Paul Yes, it's a bi-directional thing. If you can search for it, you can copy-paste it. If you cannot search for it, you cannot copy-paste it. –  Moriarty May 18 at 16:45
1  
It is possible to "repair" a PDF that is not properly searchable, for example by running it through OCR. Acrobat's own OCR won't run on a non-image PDF, but some other OCR packages will. –  EnergyNumbers May 19 at 14:09

If your paper has been posted to the arXiv, go to the paper's page (here's one of mine), click on "Other formats" and "download source". Assuming the authors submitted the LaTeX Source to the arXiv, you'll get a tarball with the LaTeX source, which should be a lot more searchable. (Yes, I do this for large enough papers where I expect to do a lot of searching.)

share|improve this answer
    
Interesting solution –  Jonathan Landrum May 19 at 18:08

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.