Hello, LibraryThing. I'm not sure if there is a better place to post/raise this issue, but I've noticed a couple errors on LibraryThing that relate to proper alphabetizing/filing rules:
1) According to the American Library Association's 'letter-by-letter' rule:
http://www.ala.org/tools/libfactsheets/alalibraryfactsheet27
authors should be filed/alphabetized letter by letter. Adhering to this rule would look like this:
Oates, Joyce Carol
O'Connor, Flannery
That is to say, names with prefixes (such as 'O'Connor') should be treated as one word and alphabetized letter by letter. So, Flannery O'Connor would be alphabetized after Joyce Carol Oates because 'Oa' comes before 'Oc.' Anyway, I noticed this rule is not currently being followed on LibraryThing, as Flannery O'Connor is presently alphabetized before Joyce Carol Oates. Any chance of fixing this?
2) Carlos Ruiz Zafón, author of The Shadow of the Wind (read this book if you haven't, it's phenomenal!), is currently alphabetized on LibraryThing under 'Z' for Zafón, which is incorrect. Just like Gabriel García Márquez, who is correctly alphabetized under 'G' for García Márquez, Ruiz Zafón should be properly alphabetized under 'R' for Ruiz Zafón. This is confirmed on the copyright page of his works:
Ruiz Zafón, Carlos
Anyway, if there is a better way to alert the folks at LibraryThing of these issues, please let me know. I just thought they were worth bringing up in order to improve the accuracy of the site.

Why should LT follow the ALA alphabetizing rules? It isn't actually a library.

>1 erbisoeul: Where on LT are things alphabetized wrong? In your catalog?

You'll find that most libraries don't follow the ALA filing rules anymore, because computers don't file like that and most of us have computer-based catalogues. The ALA rules are for CARD catalogues. I stopped worrying about those the day I emptied the card catalogue drawers over the recycle bin (o frabjous day!)
Regarding Carlos Ruiz Zafron - how his name is filed depends on how it was entered. If the users entering those books entered it as Zafron, Carlos Ruiz, then that's how it will show. If you look at the Common Knowledge for his name, the canonical (official) version of his name has already been entered as Ruiz Zafron, and that's all that can be done to correct the error.

>2 jjwilson61: Because LibraryThing is a social cataloging system, which is used by, among others, libraries. To be an effective tool, it would make sense for the site to adhere to standard rules of cataloging.

>4 tardis:If the users entering those books entered it as Zafron, Carlos Ruiz, then that's how it will show.
Except that users probably didn't enter it directly but got it as part of a book record from some source. If you used Amazon as the source and it gave you bad data then you should just fix it in your catalog.

>5 erbisoeul: But most users are individuals who understand how computers alphabetize things. To use the ALA rules would probably cause more confusion than it helps.

>6 jjwilson61: I forget about the bad Amazon data thing because I don't use Amazon data myself - if I can't find it in Overcat or a library, I use manual entry. But the fact remains, the user who imported the data accepted it as-is rather than editing it to correct it. Most probably through ignorance - if you're used to standard names in most of the ENglish-speaking world, then Ruiz looks like a middle name.

Alphabetising by computer rules is straighforward and easy to understand. Non-librarians have always had trouble with ALA filing rules. Besides, this is not just a site for US members, and why should those members have to learn that way of doing things?

>9 MarthaJeanne: Plus, I expect that the programming challenges of making a computer use ALA filing rules are not worth the time it would take. And any time a member exported their data to Excel and sorted it, they'd lose the ALA sort again anyway.
At least LT drops 'The,' 'A,' and 'An' when it sorts.

Except on the combining pages now.
Totally agreed on the programming challenges. The nonstandard characters are still a hassle.

>11 MarthaJeanne: Really? Guess I haven't done any combining for a week or two. Hope they get that fixed soon!

>12 tardis: Nope. It was changed on purpose because the old code wasn't working and it was too much trouble to fix.

>13 jjwilson61: Dang!!

Tim asked and no one objected. The thread was in the Combiners group.

>4 tardis: If you follow the link I provided above, you will see that the ALA Filing rules are 'rules for the arrangement of bibliographic records whether displayed in card, book, or online format. These rules are 'letter-by-letter' (or 'character by character') rules. They also largely ignore distinctions among different punctuation marks and do not distinguish among the types of access points.' So, stating 'The ALA rules are for CARD catalogues' is erroneous, as you are overlooking their purpose as a whole. As far as your claim that 'most libraries don't follow the ALA filing rules anymore,' well, as a librarian myself, I admit you raised an eyebrow with that one.

>7 jjwilson61: Fair enough, but 'How computers alphabetize things' is technically not alphabetizing, which is what I have pointed out above. An alphabet is comprised of letters, not punctuation, which is why the ALA rules rightly ignore punctuation when alphabetizing. I admit I'm no expert on computer programming, but I find it hard to believe a little detail such as this couldn't be fixed by someone with the proper know-how. Also, properly alphabetizing words is hardly confusing.

>9 MarthaJeanne: I wasn't implying that LibraryThing was a site just for US members, although I am curious how other English-speaking countries alphabetize. Other than spelling variations, I can't see how it would be any different. I've looked around online and haven't been able to find an alternative set of rules. If you know of any, I would greatly appreciate a link.

>17 erbisoeul: I spent part of my career designing filing systems. There are without a doubt different ways of alphabetizing things, some of which are not compatible with the way computers 'alphabetize,' and some of which appear to be different from the ALA. There isn't a right or wrong way so much as there are conventions that need to be agreed upon in particular systems. If the clerk to the left files McGxxx before MacGxxx but the clerk on the right does the opposite (both ways being technically correct depending on the system) retrieval is going to be a beast. That's why a lot of filing systems use numbers, which can be sorted without ambiguity.
After all, there would be no need to have standards or guidelines if everyone did it the same way to begin with.

>18 erbisoeul: Why do you want to limit it to English-speaking?

If you're using a card catalogue or looking along a shelf, you need to know whether Oates comes before O'Connor, otherwise you waste time or risk not finding what you're looking for. If you're using a computer, you don't care, because you're never going to look for a known author name by browsing a list: you use the search tool, and all you care about is that the search tool doesn't freak out when you type in an apostrophe. Arguing about which set of rules we should use for alphabetical sorting is a bit like discussing how far apart the stables should be on the freeway. It's an interesting discussion, and it is something it would be nice to get right, but it isn't actually very important.

>20 MarthaJeanne: What was it that erbisoeul said that makes you think that?

Besides, I just looked at your link, and there are 2 different sets of rules, that give different orders! Reading it carefully, it says that you need to set filing rules, and that they can differ between libraries.

>22 Collectorator: See >18 erbisoeul:, where erbisoeul mentioned wondering about exactly that.

>24 jjmcgaffey:, yes that is the only mention of it I find. Must be a red herring.

>21 thorold: that's not always the case. If I'm unsure of the complete spelling but have some indication of at least the start than I'll peruse a list looking for the right item.
But I'm not sure that changes your point significantly.

Erbisoeul, your suggestion that LT should adopt an English-language character set for alphabetizing in order to 'fix' things does not take into account the international orientation of this site. Many users here have catalogued items from non-English languages, some of which are diacritic-heavy, and some of which contain (many) more characters than the 26 of English. Rules for alphabetizing are different in different languages, too, so alphabetizing by 'letter' is actually rather difficult to implement. (I'm ignoring computer-based alphabetizing for now).
I'll give you an example. Just for my catalogue, author names (and titles and tags etc) have French characters (é, è, ê, à, etc), Dutch characters (ë, ï), German characters (ü, ä etc), Swedish characters (å, ä, ö), Danish characters (æ, ø, å), Icelandic characters (ð and þ, etc) and Irish characters (ó, á, etc). The link you provided didn't give any information on handling diacritics (which is unacceptable) and I think it's highly likely that the American system you suggested would collapse ô, ö, ó and ò with o, and à, á, ä and å with a -- what else could they mean by 'letter-by-letter'? For some languages that is fine; but for others it most definitely is not.
Lt already has its issues with alphabetizing non-English characters (I'm ignoring non-latin alphabets for the moment, and non-alphabetic writing systems): it ignores the distinctions between ó and ö, for instance, and the alphabetizing is wonky because of that. Again using my catalogue as an example, the authors Ó Donnchada and Östergren should not be listed between A and B.
I do not think that adopting an English-language set of rules would 'fix' things; it would positively make matters much worse for many users. Remember that LT is not an American-only website -- or even an English-language only website, and the system must be able to accommodate much more than the ALA's letter-by-letter rules allow for. Restricting ALA-type rules to just the .com version of LT would not be a solution: many English-speaking users read multiple languages; and many international LTers use the .com version of the site, even when a version in their language is available.
Finally, when you say in >18 erbisoeul: 'I wasn't implying that LibraryThing was a site just for US members' -- maybe not, but when you suggest that an American set of alphabetizing rules should be followed and that they would 'fix' things, you are not only assuming (and therefore implying) that rules for alphabetizing English should be the norm, but even that they would be useful or even acceptable to all users of this site. And that is just not the case.

>27 Petroglyph: Sometimes I wish for a Like button here on LT. Hear hear!

The solution for non-English characters is the Unicode default sort, which should be easy for LT programmers to install. This would sort accents as secondary weights, only making a difference if the words are otherwise the same.
It's possible they can configure a custom sort, but anything smarter would be a lot more confusing for anyone not a serious polyglot; ñ is its own character, but ç won't be. å might its own character at the end of the alphabet, but ä can't be, because its biggest user is German, which doesn't sort it that way. (Or we don't put å at the end, because that would be more consistent?) It would be impossible for anyone without several languages under their belt to figure out how to sort a title, and a lot of languages would still not sort right.

I'm with 27>

I'm not entirely sure, but I don't really see why it would be necessary to adopt any kind of library-approved rule at all? I'm not talking about diacritics, umlauts and the like (and oh, how I would love a solution for sorting in my library!), but the given example, Ruiz Zafon - if in your library, it's sorted with Z, you'll have to change the your own author entry from Zafon, Carlos Ruiz to Ruiz Zafon, Carlos. I really can't see where else we would need authors being listed alphabetically? The LT entry lists him as Ruiz Zafon, so the 'error' is in your own library.
As for titles, I'm not paying too much attention to alphabetical listings on combine pages anyways. There are too many and way too many junk entries to have them all listed in a 'scientific' way. This is a site living by user entries, after all.
But I do want to add my voice in suggesting that we need a way of dealing with disacritis! Or maybe there could be an option of moving authors/titles inside one's own library. I have way too many Kästner's and Krüss's showing up all over my library to not care about that issue!

Of course the ultimate solution is to tag all your books with a sorting key and then sort by tags. (I'm not entirely unserious about this. My shelves are once in a while sorted by author and title, and I'm not really fond of the idea of disregarding a, an and the when sorting, so I'd prefer a more 'mechanical' sorting than most.)
If you want to sort on several fields this gets very messy very quickly, so it is not a feasible solution.
Another more technical solution is to export your data to a spreadsheet and then devise your own sorting. I do something like that only using a database rather than a spreadsheet.
Most of the time I just search to find a specific book and then I don't care as much about the sort order as long as it is approximately what I'd expect.
ETA: I'm adding some of Sue Graftons books and of course 'A for Alibi' sorts as 'for Alibi' which looks silly.

>16 erbisoeul: Well, I just checked, and our online catalogue does file by ALA rules, so your raised eyebrow is deserved :) In my defense (feeble though it is) the ILS software handles all of that so I don't have to think about it, and in my library we don't file by author (LC# plus cutter for main entry). I'm not sure if users often browse the catalogue by author, so I doubt they'd notice one way or the other.

>32 bnielsen: You do know about the double pipe for adjusting the sort? If you enter the title as ||A for Alibi, it will sort as 'A for Alibi' instead of 'for Alibi, A'. Similarly, if you have an article that LT doesn't recognize at the beginning of the title (or whatever), put the double pipe in front of what you want the title to sort by and it will. Doesn't help with diacritics etc, but if your problem is that it's sorting by the wrong word the double pipes are a lifesaver.

>34 jjmcgaffey: Thanks. I do know about the double pipe, but I'll live with funny sorting rather than pollute the title with pipes. BTW some of the library records also contain stuff like that, a is supposed to mean aa sorted like å, I think.
To me that's stuffing two things into one field and I try to avoid doing that.
But yes, it's a way of fixing 'A for Alibi' or 'A is for Alibi' if you have the English version. Actually LT could have implemented sorting so it would notice that 'A for Alibi' is a Danish edition, so eliminating A in the sort should not apply. But no doubt this would make LT more complicated and slower. I prefer simple rules, so I'll just live with the quirks of the current regime :-)

I have found the above discussion very interesting and have learnt a few new thing, eg. ALA 'letter by letter rule'.
Just wondering where Üni-Code' fits into this..?

>36 guido47: New readers can start here (Hint: You'll need lots of coffee or stronger for this).
http://www.unicode.org/reports/tr10/

I looked just a bit, and the result is (as we have known for a long time) there just isn't a way to sort that will please everyone. For example the character ö can be sorted, depending on language and context:
as o (usual in English and other languages that don't use it)
as oe (German for libraries, etc.)
after oz, i.e. next letter after o (German for phone books, etc.)
as one of several letters that come after z (Swedish)

Thanks >37 bnielsen:,
As a Once time programmer when I read Collation order is not a stable sort.
Stability is a property of a sort algorithm, not of a collation sequence. I realized I would indeed need a stiff drink. Not just coffee.

>38 MarthaJeanne:as oe (German for libraries, etc.)
Interesting. I didn't know that. Turkish seems to sort like German.
as one of several letters that come after z (Swedish)
Yes, from A-Z is from A-Ö in Swedish.

>35 bnielsen: The double pipe only shows up when the title is edited, not when it's displayed, so it shouldn't 'pollute' the title.

Maybe I did it wrong, but my double pipes showed up the day after I put them in, and they still show. Not in my catalogue, but on the main book page.
http://www.librarything.com/work/605934/book/29067462

They aren't supposed to. You should report it as a bug.

I think the double pipes are also stoved into the tab-export and I don't like stove pipes there :-)
>39 guido47: At least I hope I provided able warning so you had the ingredients at hand.

>44 bnielsen:
Speaking of the tab export, did your bug ever get fixed?

>45 lorax:. The bug is with the Marc export, and the answer is no, but it has been very well defined. Some 180 of my books are just missing in Marc export. All of them have been added with 'Det kongelige Bibliotek' as source so that's a smoking gun.
The protocol for fetching the Marc export is a bit weird, but I've finally gotten around to scripting it.
..
Columns: 33/(66242), Rows: 2/6201, rdbtable ok: lthing.rdb
- - checking 6201 books -- 922 tagged as recycled - - 6019 in marc format - -
- - checking 6106 books with cover artist descriptions - -
So it's 182 books missing as of this moment.

Rousing discussion everyone. I'm glad I could bring ALA rules to the attention of anyone who was unaware of them, and thanks to those who contributed to this discussion in a meaningful way. I agree that there is no method of cataloging that will please everyone, but that doesn't mean the issue isn't worthy of discussion. Just to be clear, I am an American librarian and I, of course, realize that these rules do not apply to all languages and was not suggesting otherwise.
>27 Petroglyph:'your suggestion that LT should adopt an English-language character set for alphabetizing in order to 'fix' things does not take into account the international orientation of this site.'
Most users, including myself, are aware that LibraryThing is comprised of multiple language sites, such as ara.librarything.com (Arabic) and jp.librarything.com (Japan), etc. When I suggested the implementation of ALA rules here on LibraryThing, I assumed individuals would logically deduce that I was referring to the main English site at www.librarything.com Apologies for not spelling this out in my initial post, but I figured it was self-evident.
'Rules for alphabetizing are different in different languages, too, so alphabetizing by 'letter' is actually rather difficult to implement.'
Transliteration, i.e. not difficult at all.
'The link you provided didn't give any information on handling diacritics (which is unacceptable)'
Diacritics are ignored, same as punctuation. They are also not exclusive to a particular nation as you suggest.
'Restricting ALA-type rules to just the .com version of LT would not be a solution: many English-speaking users read multiple languages; and many international LTers use the .com version of the site, even when a version in their language is available.'
When in Rome, do as the Romans do.
'..when you suggest that an American set of alphabetizing rules should be followed and that they would 'fix' things, you are not only assuming (and therefore implying) that rules for alphabetizing English should be the norm,'
Actually, it's a suggestion, not an assumption. I suggest ALA rules because they properly alphabetize the Latin alphabet, i.e. letter-by-letter, and ignore characters that are not part of the alphabet. As I have pointed out in my initial post, the implementation of punctuation, such as the apostrophe in 'O'Connor,' leads to errors in proper alphabetization. Furthermore, I did ask above if there were other ways of alphabetizing in English and, so far, no one has pointed to a concrete alternative (see 18).
'..but even that they would be useful or even acceptable to all users of this site. And that is just not the case.'
No kidding and, again, I never suggested this.

The different language sites only show different translations of the text on the site pages. It doesn't affect the actual way the site works, like sorting, at all.

>48 jjwilson61: I noticed this as well, at least in the French, Pirate, and Japanese versions I took a look at. However, surely there must be differences in sorting, at least in so far as native users are concerned. My own library remains unchanged when I switch to the Japanese site, as it was entered using the Latin alphabet and transliteration evidently does not occur. However, if someone from Japan is using the site and entering data in their own language (assuming this is possible), I wonder wouldn't there be differences in how that language is cataloged?

No - if you entered a couple Japanese books (in Japanese), they would be sorted in your library exactly the way they'd be sorted in a library that was all in Japanese. The same sorting rules apply across all the sites. Which means that the sorting rules on LT are extremely complicated and complex..which is what we've been trying to say.

>47 erbisoeul:Transliteration, i.e. not difficult at all.
That's horrible. I've never seen an alphabetized index or anything that sorted things by their transliterated forms, and I never hope to see one. Even in something like Russian, it could be quite frustrating to look for a Russian book, especially if you don't know exactly what transliteration is being used. Georgian? Converting თ to t and ტ to t', and then sorting them by ignoring apostrophes are going to be miserable.
The only reasonable way, IMO, to sort material in multiple scripts is to put the scripts in some order and sort them internally according to their own rules.
Diacritics are ignored, same as punctuation.
What about eth and thorn? What about the the other hundred or so non-diacritic characters in the Latin script?

-snip-

barf bowl

We want dd, ff and ll digraphs for Welsh, or we'll blow up yr mailboxes!

I knew it was only a matter of time before Meibion Glyndŵr found their way here.

And this after the Dutch convinced Tim that their digraphs should always be read as two letters no matter how they had been coded.

>56 MarthaJeanne:: Yes, because that's how our digraph ij should be read. You will find it under 'i+j' in any Dutch lexicon, even though it's sometimes - mistakenly - used as 'y'.

>57 Nicole_VanK: But in encyclopaedias and indexes it is alphabetized as 'y'. Yes, we are not very consistent.

>58 henkl: I myself would have no idea how to sort on all our weird letters. I admit I would sort IJ at Y, and OE at O, EU at E etc. No idea of the correct sort order.

I am a literal freak, I love reading this thread. :D
A long time ago, in a public school system in the US, we were taught alphabetizing (if that isn't a word, it should be, ha!) that apparently differed from what is being discussed here. Names such as O'Connor and O'Reilly were at the beginning of the O section, followed by O entries without an apostrophe. The same system was used for Mc and Mac prefixes. I thought it was used in our library, I'm fairly certain of it. This system was also used in phone books for many years, and might still if there are any phone books still being published and distributed.
Anyone else recall this system? I can't be the only one..

Yes, I've seen that system used in indices.

>60 fuzzi: Yep

Somewhere I was taught that 'Mc' names should be alphabetized as if they were spelled 'Mac'. Not sure what system that is..

>60 fuzzi: That's what I remember, too. And not just from school. In the early 90s, I was a file clerk and that's how I had to alphabetize.

The Mc/Mac rule was in use in libraries back when we were typing all the cards, and then alphabetizing all of them by hand. I think the decision was made somewhat arbitrarily, but in library classes we were taught to follow it. Then, when computers made their appearance, we found that they alphabetized letter by letter automatically and we had to change the way we did things.

All alphabetizing systems are arbitrary. All that matters is that they are internally consistent.

Ugh, when I started as a page in the library (wayyyyy back when) I got given the 'page' test. Where of course you have to put the books in the correct order. So I do, putting both Mc and Mac where they should go in a regular alphabet. Then I get told (to be fair my to-be boss was smiling at the time) that I only got one wrong on the test.
For some odd reason, at that time in that library we were filing both the Mc and the Mac authors... before even the Maa, or Mab authors.. *doh* :)

>67 DanieXJ: yes, that's the way it was done, apparently 'bc' (before computers).

>67 DanieXJ: That's how it was done when I was a file manager & records management specialist in the '70s & '80s.

>67 DanieXJ: Your post made me smile, because my son became a library page a few months ago, and we were teasing him because he had to take the 'page test'.
When I was a child, it always bothered me that Mc and Mac were shelved together.

I worked as a temporary secretary for a few years, and the question of 'which filing system' always needed to be asked early. O' first, or treated as O(space) (which mostly put it at the beginning, but not always), or treated as O(next letter)? Mc and Mac together, and if so with Mac or at the beginning or at the end? Did 'the' count in a name/title, or did it get ignored? I swear every office I went to (dozens!) had a slightly different arrangement of files.
I never worked in a library, but I've looked for books in plenty of different ones, and had the same 'where do _these_ people put them?' question.

I think the problem with many of the IMHO over-fancy schemes is that they only make sense to a select few. The Danish authorized filing rule used in dictionaries is that 'aa' sorts as 'å' (i.e. the end of the alphabet) if it is pronounced as one vowel, but as 'aa' if pronounced as two vowel.
The examples where this makes an actual difference are few and far between, but I've looked for the word 'kraal' (South-african origin, I think) and couldn't find it in the dictionary because it was filed at the bottom of the kr* words.
Hmm, maybe I should write a book and change my name to Kraal just to see where the librarians would put it. :-)
But it's possible they have another rule for author names.
My vote goes to simple rules. If a publisher calls an author McArthur on one book, and MacArthur on the next, I can live with looking for his books on two different shelves if need be. Even Evtjuchenko, Jevtjujenko, Yevtjuchenko would be ok.

Can you put the Title's of the books in alphabetical order. I did the Authors but would like to see the titles that way. i'm new here and just starting a small library . I thought it would make it easier for patrons that view my library online to find books if they were in alphabetical order. Any suggestions???
thanks

>73 Beckysbooknook: Just click on the word 'Title' at the top of that column. You can do that with most columns, and your visitors should be able to do the same thing.
If you want to sort by two things, say, by author and then by title, look for the icon with the one-up, one-down arrows, near the printer icon, above the catalog columns. You can sort in all different combinations.

Note that your patrons will have to sort for themselves. You can suggest a style, but that does not include the sort.
BTW when you change the sort by clicking on a column, the previous sort becomes the subsort, so you can aslso sort by author and then by title by clicking first on title and then on author.

At the bottom of a list or cover view is a Permalink. If you copied this, would it include the sorting preferences?
James

>76 Keeline: Yes, and it even seems to work now. It's been rather wonky.. It doesn't do subsorts (or rather, it has two sorts but they're always the same - I tried modifying the URL, didn't work either. The last one is what it sorts by).
>73 Beckysbooknook: So if you sort by title, then go down to the bottom of the page and copy the URL of the permalink and give that out, your patrons will see your collection sorted by title. They can then sort other ways, if they want, but as long as they use that URL to get to it it will start out sorted by title.

@73-77: I once wrote up an unofficial wiki page to document what can be encoded into the URL/permalink. Beckysbooknook, you might find it useful for constructing a link to your library.