But paper abstracts are always free (even with paid journals).

If you have to scan random papers to find a solution to a problem you cannot name, I fear for your well-being :-)

> If you have to scan random papers to find a solution to a problem you cannot name, I fear for your well-being :-)

The disadvantage to being a generalist is that I don't know the terminology of the field I need to research. Take the texsearch example above. Given a LaTeX string I want to find similar strings from a large corpus so I start by googling things like "code search" and "syntax tree search". After scanning a few dozen papers and following links I find that the magic search term is "approximate string matching" which nets me an overview paper. I scan through the links from that paper and dismiss most of the algorithms as unsuitable for my particular problem until I'm left with a few candidates for prototypes.

Back then I had access through my university. Today I wouldn't be able to read half of those papers. For recent math/CS papers I can usually find a preprint but anything else is a struggle.

OK, so how did you do your research in the end? Do you think you missed much because of not being to access certain papers? I mean, it seems that you found a way to access at least that paper, or did you pay for a subscription to a certain journal?

At the time I had access through my university. Since leaving research has become much more time-consuming. My local library doesn't have any subscriptions. If I really need to find something I can take a two hour round-trip to the British Library, although its opening hours align neatly with office hours. In an age where the worlds information is at our fingertips its ridiculous that I have to take time away from work and get on a bus just to do basic research.