A big problem is that these pdfs are images rather than OCR'd text. Gutenberg is such a useful resource as the "full texts" of public domain _title_s (including Huck Finn) are available. By that I mean the text is available to look at, read, copy and edit.

If I download a PDF from google, all I have is a series of images on my machine. This is not useful to me. Also, some of the older editions could benefit from being OCR'd simply because there is too much dirt and mess on the original scan. I'm interested in the preservation of the word, not the image behind it.

If you can force yourself to scan through google books, it's full of shit. A lot of books are such twaddle that they are just filling up space and no-one wants or needs them. Searching for popular 18th-19th century _title_s I had to wade through piles of crap to often find that while there were plenty of books and magazine referencing them the _title_s I wanted weren't there themselves.

If I want classics I will go to Gutenberg myself. I can easily convert between the many formats that texts are hosted under there, so if I really need that pdf file I can convert to it.