FWIW, sentence case is ugly, but susceptible to programmatic case changing. A few lines of [language of choice] can massage "Harry Potter and the potter wheel of doom" to "Harry Potter and the Potter Wheel of Doom" - but the other way around is a clever trick indeed.

Now will someone tell me why so many thesis authors are still told to enter their titles in all caps? Oy.

I hear you Thomas (and I would use Perl, FWIW), but I still wonder what possible purpose sentence case serves over the way in which it appears on the item itself. Why do we go to the trouble to change it in our transcription, then have it display oddly to our users, and be forced to change it back to title case when we want it to look right? I still don't get it. Someone please help me understand.

Karen's question on this had some interesting responses on rda-l back in January this year. The debate apparently goes back (at least) to the mid 19th c.

One problem with transcribing capitalisation of titles from the item is that they are often in all-caps, eg DE ANIMA (ON THE SOUL).

Personally, I've grown to like sentence case. It makes it easy to distinguish some things in a title, for instance: "A discourse on The meditations" might let you know that this is a book that is a discourse on another title, The meditations. And, as Thomas Dowling says, it is easier to go to title-case then the other way round.

But I've always suspected that the real reason behind this is a latent francophilia.

Relatedly why does APA style do the same but MLA doesn't?

I'm having trouble linking to the rda-l discussion but this quote from John F Myers summarises most of the arguments for sentence case:

"Why we would prefer sentence case over title case is debatable. I can
think of several factors. First, title case requires making judgments
as to which words should be capitalized, a process that takes time and
that distracts the cataloger from the other important details of the
record. Second, when cataloging was comprised of typed catalog cards,
minimizing the need for capital letters reduced the complexity of key
combinations and the incidence of double capital letter typos. Lastly,
sentence case makes it clearer that an embedded title is present.
Cutter and ALA are quite clear in capitalizing the first word of a title
that appears within another title. In days of typed or even
hand-written cards, offsetting an embedded title by italics or
underlining would have been quite challenging. Normalizing titles
through sentence case would have been the most viable alternative.
(Post-scriptum, can any of us imagine the nit-picking debates between
catalogers over disagreements as to application of title case?)"

Martha Yee also made the point that sentence case allows you to distinguish when a word is used as a proper noun, eg 'The road to Perdition'.

Like Erik, I've grown to like sentence case but having worked for a while now with non-cataloguers who create metadata, I realise that title case is how they naturally want to enter titles.

amen! "Sentence case" makes no sense to a non-librarian (and maybe not even to librarians-who-aren't catalogers) Aren't we cataloging for the user? Shouldn't we do what makes sense for them? <sigh>

I think that a lot of these arguments can be boiled down to two points:

(1) Sentence case isn't "lossy." Use sentence case and you don't know which words are proper nouns and which aren't.
(2) Sentence case requires subjective decisions.

#1 is why you can go from sentence case to title case easily. But to do the reverse you'd need deep knowledge of the characters, etc.

#2 actually grows out of #1. Something like "Road to Perdition" actually trades on the lossy nature of cataloging rules. You don't know Perdition is a place until you read the book, and, I suspect, that's exactly the way the author intended it. By capitalizing Perdition in a sentence-case context, the reader knows something the author—unless a librarian—didn't want the reader to know.

You know what's lossy? This blog! I used paragraphs, but this blog makes me look like a rambler... :)

The wikipedia article on Capitalization does a good job on this, particularly the section #Sentence_case_versus_title_case. BTW, I've noticed the Teach Y(y)ourself books currently use all lower case in their titles, ee cummings-style eg teach yourself beginner's japanese script.

This has all been terribly informative and interesting. But as we try to move up the data stream and use more publisher-supplied metadata to feed our cataloging systems, I wonder how much staff time we will want to invest in sentence-casing titles. This is because publishers almost uniformly use title case. For example, see roytennant.com/proto/onix/onix.cgi?field=&query=library .

Yes, I imagine worrying about uniformity of capitalisation will be the first thing to go when publisher and user-supplied metadata feed cataloguing systems, closely followed by punctuation.

Thesis titles in all caps, scientific works with equations, incunabula, etc. are all somewhat problematic when it comes to transcription. The rule wasn' tcreated out of thin air -- there was a reason at the time. The time hasn't completely passed, either.

Actually APA is inconsistent on sentence case. It uses sentence case for citing books as books, but upper case ("book title case") when the book title is used to cite a chapter. Catalogs are now displaying a mixture of sentence case and title case and ALL UPPER CASE, with vendor-supplied order records being loaded directly. Frankly, I think this makes the catalog look as though there is no one in control of overall data quality, which also creates an impression on end users, probably not a good impression. Whatever the standard, I suspect that consistency is something that catalog users have come to expect from us, just as they expect this consistency from the APA, MLA, Turabian, or Chicago style manuals.

many users, such as myself, don't bother to capitalize. perhaps it's time to throw away capitalization? it also has the advantage of making certain words more neutral.