What's the title of the CD track you're listening to right now? Either you know it by heart, or you have to get the program liner from the box. Think about that. You have a digital storage medium which can hold over 700 MB of data, and the most important information is stored on a paper liner in its box.
Why can't the CD format contain the title of the CD, the titles of the tracks, and maybe the entire booklet? The standard for a CD doesn't even require an identifier to be stored on the CD! If I'm reading a book (you know, the big bound bits of dead trees), I can look at the cover and discover its ISBN. If I'm listening to a CD play, my player has no idea which CD it is!
Any computerised CD player program could use all of this data. It could display the CD name and track name, it could recognise the CD and use your favourite track list, it could automatically catalogue all your CDs, it could do a million things which would make it a more exciting, useful program. And it would cost the manufacturers $0 to do it. Expensive (non-computerised) CD players could make use of the information. Cheap CD players could ignore it. It would cost $0 on to add it to a cheap player, because nothing would get added!
You could do it in Unicode (to allow for non-English CDs), and have the cheap programs ignore all non-European alphabets. And it would still be better than nothing.
Of course, some CD player programs do identify the CD being played. But since the ID number imprinted on (some) CDs is both optional and nonunique, they don't bother with it. Instead, they look at the track lengths, and compare that against the track lengths stored in their databases. This is error-prone, this is slow, but most of all, this is moronic. Why couldn't a compulsory ISBN-like ID number be part of the standard?
Having identified your CD, the good player programs are gracious enough to allow you to type in the title and track titles, so they can display them. This information was known to the manufacturers; they even supplied it to you! On a slip of paper, of course; was there anything else in the box that you could put data on?
Some really cool programs use CDDB (or, better nowadays, FreeDB) to download this metadata from the Internet. They measure the track lengths, then ask a server if it has titles for this CD. In return, they (sometimes) get back the requested data, as typed in by someone else. Maybe there were mistypings. Maybe the CD was misidentified. But why do I have to go get the (possibly damaged) data from some anonymous third party, when the first party (the manufacturer) had it in perfect form?!
Not storing any text metadata for CDs is probably their biggest flaws. And it doesn't seem like it's going to be fixed. Even if they did fix it today, it wouldn't solve the problem for the old CDs. drinkypoo informs me that CD TEXT is a belated attempt to do just this; predictably, it fails.
This rant is a product of discussions with haggai, who probably feels the same, but might not use the word moronic.