Ebook reader that can handle large books and search them quickly

[pruss]

I started in another thread by looking for an Android reader that supports display scrolling rather than paging, but I am now thinking that's too picky. I have three feature must-haves:
1. open large files quickly for quick reference;
2. search large files at a decent speed; and
3. handle intra-document links.

At this point I'm not finding anything satisfactory. I am blogging my on-going test results here. So far I've tested Moon+, FBReader, CoolReader, Aldiko, Nook app, StarBooks, Foliant Beta, Mantano Trial and Kindle app.

I've tested all of them with the same Calibre-made epub file generated from Aquinas' Summa (25mb of html, broken into 626 files, generated from a Plucker file), except for the Kindle app where I used Calibre to generate a Mobipocket file instead.

Moon+ and FBReader are the only readers that I found to have a decent search speed (I didn't test Coolreader's search, as Coolreader was otherwise unacceptable), namely they all can search the file in about twenty seconds, which is about as good as can be expected without pre-indexing. (There is nothing amazing about 20 seconds: that's how long it takes to unzip and grep the epub from a shell on the device.)

Foliant Beta is marginal at 35 seconds. The other readers either have no working search (Starbooks) or take somewhere between 74 seconds (Aldiko) and 6 minutes (Kindle) to search.

Aldiko, Nook, Starbooks, Foliant Beta, Mantano and Kindle all load the document very fast (some of them only after an initial slow import, but I can live with that). The other apps take anywhere from 20 seconds (Moon+) to a minute (Coolreader) to load the test file.

I couldn't get the intra-document links working on Coolreader and Foliant Beta.

So none of the apps satisfied all three must-haves. I will eventually need to move over (up? down?) from Plucker on my PalmTX. For comparison, the Plucker version of the file loads in about two seconds on my PalmTX even with the Palm underclocked to 208mhz, while my Android device is an Archos43 with a 1000mhz CPU. The Plucker version takes about 42 seconds to search, which is too much, but understandable with a slower CPU, and still faster than Aldiko, Nook and Kindle on the faster CPU.

There are many reader apps I haven't tested. I went with what people recommended in the other thread, or stuff that looked popular on the Market. So I will be happy to try others, if they look hopeful.

Besides the three must-haves, there are some other things I'd like: scrolling instead of paging, open source, no margins, small margins or adjustable margins (the Nook app showed the document with absurdly large margins).

[pruss]

My search for an ebook reader that can handle large interlinked books continues. Now I'm up to 11 readers tested, none fully satisfactory, with Moon+ still being the best (though not quite good enough):

Wordoholic: Failed to load the Summa epub at all: just gave an error message.

Captionary: Loaded in about five seconds, but links don't work and no search.

[pruss]

If anyone wants to do the tests themselves, here's a somewhat more up-to-date version of the Summa epub. (So far, this version has only been tried on Moon+. Loads in about 25 seconds the first time and 18 seconds the second time. Better. I suppose omitting the cover sped things up a little.)

[=X=]

Originally Posted by pruss

If anyone wants to do the tests themselves, here's a somewhat more up-to-date version of the Summa epub. (So far, this version has only been tried on Moon+. Loads in about 25 seconds the first time and 18 seconds the second time. Better. I suppose omitting the cover sped things up a little.)

The link is not working

[pruss]

Sorry, link fixed above, and it's also here.

[pruss]

Playing with the FBReader and tweaking it, I see the problem. FBReader insists on reading and parsing the entire epub file into memory at once. This means that once you've loaded the file, going between different parts of it is really fast. But the loading takes too long. Plus the memory usage will make even larger epub files completely unusable. I have Plucker files that are much, much bigger than my test file.

[pruss]

If anybody is following my question, I found a solution to the problem of *reading* large documents. Use Microsoft's HTML Help tools to generate a chm file, and then Moon+ reads it instantly. Very nice. However, Moon+ then only searches the current html segment (I reported this to the developer--I think he'll look into the large file issues eventually).

Searching in chm could be really fast as chm files can be compiled to include a full-text index.

[=X=]

Im pretty much getting the same results.

[pruss]

The Moon+ developer asked me for my test file, and I think he's going to work on the issue.