ISBNs and Amazon Japan

I don't know where to post bugs anymore. Sorry if this is the wrong place.

The ISBN problem with Amazon.co.jp is a bug of long standing that's been reported quite a few times.

When you search for a book on Amazon.co.jp and add that book to your catalog, the ISBN does NOT transfer into your data. Instead, you have to pull up the edit screen for each book, then copy or type in each ISBN manually.

Without the ISBN, the cover won't import either. If you have lots and lots of books to enter, this gets tiring. So very tiring.

Will it ever be fixed?

My friend Bruce added the book Musashi to his collection using the Amazon.jp database.

The Amazon entry had the book name in Japanese and English and that is how it looks in his catalog:
http://www.librarything.com/catalog.php?view=brucedavey&searchmode=Books&amp...

I noticed that when I clicked on his profile that Japanese characters didn't show up in 'Random books from brucedavey's library' feed. Only the English-language words showed.

When you search for a book on Amazon.co.jp and add that book to your catalog, the ISBN does NOT transfer into your data. Instead, you have to pull up the edit screen for each book, then copy or type in each ISBN manually.

Without the ISBN, the cover won't import either. If you have lots and lots of books to enter, this gets tiring. So very tiring.

I'd like to see this fixed, too. I've had to manually re-enter Japanese-title ISBNs up the wazoo... repeatedly (some of my titles have been entered or corrected as many as four times due to the recurring Unicode issue).

If we can get the ISBNs to "stick" - and ditto with the covers - I will be one ecstatic customer!

Sometimes this also happens with amazon.de. I ended up searching by ISBN (which I do anyway) and before hitting enter, I copied the ISBN just in case I had to manually enter it again on the edit page. It seems to happen less now though.

I'd like to see this fixed to. I hope it will happen as the unicode fixes start to trickle in.

Retaining the ISBN on an Amazon Japan import would be a blessing. While the batch upload process + pasting the ISBN back into the record is significantly faster than doing all the bibliographic data entry by hand, it's still time-consuming and I risk pasting the wrong ISBN into the record (DATA DISASTER!).

Okay, that moment of panic is over.

I still have hundreds of books left to enter in my catalog. If the amazon.jp ISBN fix is on the radar, I would like to wait until it works properly. However, if this is the lowest of LT's priorities, just knowing that would be helpful.

Consider this another vote in favor of fixing the Amazon Japan ISBNs! I'm using the tips people have suggested (copy the ISBN first, etc.) but it's so very tedious. Is this going to be a difficult fix?

I'm grateful for the bigger Unicode fix, btw. I wouldn't even be here entering titles if that hadn't been fixed first.

Tim? Abby? Chris? Christopher?

Feedback on this question, please?

Hey, don't forget Robyn!

OK, then:

Tim, Abby, Chris, Christopher, Robyn, and any other LibraryThing staffers that I may have inadvertently missed out... Can you please give us some feedback on this question?

If you say that this bug fix is not at the very top of your To Do list, we will certainly understand. We do appreciate how much you've got on your plate.

However, we would be very grateful to be told whether this is even ON the list... and if so, approximately when it might finally be tackled.

In fact, any sort of reply would be most welcome.

Thanks in advance! :-)

It would be both a site improvement and a life improvement.

I've "starred" this post, and I'll try to keep it alive until someone, anyone responds to it.

It would be both a site improvement and a life improvement.

I've "starred" this post, and I'll try to keep it alive until someone, anyone responds to it.

Me too.

This also happens on LIBRIS, the Swedish central library database. In addition, the character encoding on what's imported is broken. iso-8859-1 vs utf-8?

Maybe there needs to be a new thread dedicated to all special characters and the problems affecting them...?

I don't know if there's a common issue underlying these problems, but if so, our requests may be lost in a host of tiny individual threads (one for Japanese, one for Swedish, one for Georgian, etc.).

I don't know how or where LT announces bugs that have been fixed, so I'm regularly testing to see if the ISBNs from Amazon Japan are now importing. They are not.

Are the LT staff aware of the problem?

Updating the thread.

It had occurred to me that, since unicode-based titles are never going to be recognized by the site's software for sorting or matching purposes, we could at least get the _sharing_ part of LT among ourselves if the ISBNs were being matched rather than titles. Right now, title matching results in pure chaos; and if you separate too many of those bad combinations, the information page locks down.

So why couldn't there be an option to have our unicode titles matched for ISBN?

But, at present, not even that could work because the majority of users don't realize that they have to type in all their ISBNs themselves, so those boxes are empty blanks in their catalogs.

Edited to add: Look here, too.

I think ISBN matching is on their list of things to tackle. It would solve their matching books with no author requests to. In any case, Hooray! that they are going to look at ISBNs from amazon.jp.

It had occurred to me that, since unicode-based titles are never going to be recognized by the site's software for sorting or matching purposes, we could at least get the _sharing_ part of LT among ourselves if the ISBNs were being matched rather than titles. Right now, title matching results in pure chaos; and if you separate too many of those bad combinations, the information page locks down.

So why couldn't there be an option to have our unicode titles matched for ISBN?

Another vote for this to be implemented if it hasn't been do so already. I remember asking somthing similiar on the Google Group about work matching and ISBNs though so I think it's probably not.

(And LibraryThing is never going to recognize Unicode? Really? Gah. That's just terribly short-sighted.)

Is there a list anywhere of the site improvements that the LT staff have agreed to tackle (whether sooner or later)?

Or for that matter, a list of things that they've looked at and decided they won't do?

Yes. Tim has a long long list of both these things but it quite deliberately isn't published anywhere.

In general, I'm just wondering about the utility of posting a request over and over without getting a response. Does this simply show that we care, or is it actually counterproductive and annoying?

I honestly don't know whether in this sort of situation it's better to keep trying (squeaky wheel), or to just gracefully give up.

I could cheerfully accept either "Yes, we'll get to it eventually, but bear in mind that it's waaaay down on our to-do list" or "Sorry, we simply don't think it's important enough". (Actually, even "We've read your request and we're thinking about it" would be nice.) But no-response is frustrating because there's no way to know what it means.

I could cheerfully accept either "Yes, we'll get to it eventually, but bear in mind that it's waaaay down on our to-do list" or "Sorry, we simply don't think it's important enough". (Actually, even "We've read your request and we're thinking about it" would be nice.) But no-response is frustrating because there's no way to know what it means.

I can't second this strongly enough. I can only interpret "no response" as meaning they don't care and aren't working on it. A publicly posted list would cut down on the number of people begging for the same fix over and over again. I've even seen "to do" lists that have the number of users requesting each bug fix or improvement. The developer doesn't even have to respond as the existence of the list itself, as well as the count, reassure users that 1.) their request has been noted and recorded,and 2.) either they're not alone in thinking that something needs to be fixed, or, if they are alone, that their fix is perhaps not so pressing. I don't think Tim and his team are doing themselves or their users any favors by not letting us know that they're paying attention to these threads.

I too would really like to see some sort of poll system for logging requests. If a proposed improvement's priority level is determined in part by the number of people wanting it, this would certainly simplify things for all of us.

The other question I have is whether the ISBN issue we're facing is unique to Amazon Japan alone. Certainly I know that the Unicode problems affect many members with books in non-Latin character sets. I wonder whether we may be ghetto-izing our problem by discussing it solely under this thread.

Just as a closing comment... I've enticed quite a few acquaintances to become paying members of LibraryThing. I also have a long secondary list of friends and colleagues who share my enthusiasm for things and books Japanese. At the moment I'm putting my recommendation of LibraryThing on hold with the remaining members of this second group until (1) the ISBN issue is resolved, and (2) it looks like the character-corruption issue truly won't rear its head again. This isn't out of vindictiveness, but simply because I wouldn't wish it upon others to experience these frustrations.

On the whole, I am still pretty happy with LibraryThing. I only wish that we would get a response on this question. Given that so many people have asked about it, I too believe that it deserves an answer. It may be an issue only for a minority, but some of us within that minority own a LOT of Japanese titles!

Just spotted this under the thread IMPORTANT: New Amazon search order. Note the bolded bit:

* * * *

Message 1: timspalding
Yesterday, 2:28pm

I've received a couple complaints about it being hard to find books through Amazon. One did "spiritual autobiography" and despite their being a number of books with exactly that title, they didn't show up in the first five pages. It turns out the Amazon search request I used was tilted toward sales rank--so high-selling books jumped to the top, even if something like "spiritual autobiography" was a subject, or in the description.

{snip}

Any other Amazon complaints--besides the .jp ISBN issue, which I'm also going to try to hit--since I'm in that code?

* * * *

Seems Tim did hear our prayers (hooray! hooray!) ... thing is, he happened to respond in a different thread! :-)

(I haven't had any issues with Amazon searches per se, so it would never have occurred to me to look to other such threads for a response on our problem.)

P.S. Edited for mucked-up formatting!

(And LibraryThing is never going to recognize Unicode? Really? Gah. That's just terribly short-sighted.)

Um, I don't know that for a fact. But it does seem likely. My overall impression, based on the lack of responses, has been that support for unicode data ranks at the bottom of the priorities list.

Which is where we get into the tag workarounds and fiddling with the author displays. :P

I honestly don't know whether in this sort of situation it's better to keep trying (squeaky wheel), or to just gracefully give up.

Same.

I'd also linked (#16 above) to that "Amazon search order" thread; I'd found it discouraging that a response to what we've been asking about was hidden in an unrelated thread where no one might see it. I mean, maybe it sounds like I'm unhappy with LT and ungrateful for the fixes we've already gotten, but that's not the case. I'm just . . . honestly perplexed about how we've earned this ice-cold shoulder.

I wonder whether we may be ghetto-izing our problem by discussing it solely under this thread.

Probably so. I saw something about some other Amazon-type imports having a similar problem with ISBNs.

But if you'd meant the unicode thing in general, I'd focused on the Amazon Japan thing out of a whole constellation of issues because it seemed reasonable that this one could be fixed. Plus, it had been a long-standing, one-point source of frustration for a lot of people (who might not care about searching or information page chaos or matching or other issues), and the longer it went unaddressed, the bigger the problem ballooned as people continued to add hundreds of ISBN-free items to their catalog.

Thing is, whenever discussions or adjustments come up that affect those using unicode, they aren't finding out. At present, those affected are only meeting as mismatch collisions on the book information pages; otherwise, they're scattered over dozens of different specialized language groups. Now that the Google group has been dumped, there's no central location for the people to find out about this one topic of mutual concern.

For instance, when that fix for the mangled data when up, I'm positive the majority of people who needed to use it never knew it was there. It wasn't mentioned in the LT blog, and the LT staff didn't post about it in any of the specialized language groups. Individual users had to bring it up on LiveJournal and in the LT groups; it was passed around strictly by word of mouth, and most people weren't lucky enough to be reading the right places, at the right times. I've also posted in a number of places about working with the ISBN thing, but people are still popping up constantly who haven't figured out why their covers aren't displaying.

I'm minded just to keep on complaining, if only in the hopes that perhaps TPTB will actually fix the problems in order to make us shut up.

I'm tempted to start a new thread each day entitled "Yet another day I can't recommend LibraryThing to my friends or buy gift memberships for people because there's no Unicode support". Obnoxious, of course, but it'd be oh, so satisfying. >:D

I'm just . . . honestly perplexed about how we've earned this ice-cold shoulder.

Mikeneko, I don't understand it either. I keep thinking, did I say something that was inadvertently hurtful? I don't think so, actually, but I'm at a loss as to how to explain it otherwise.

The irony is that I've been banging the drum about the magnificence of LT for close on a year now to anyone who will listen (and many who will not); in fact, September 15 was my first anniversary as a LibraryThinger (yay!). I'm not drumming quite as loudly these days, admittedly, but that's because I want to see whether this rough patch of unclear communication will clear up. Perhaps these are growing pains. My fear is that this is becoming an institutional Achilles heel, but I will be happy to be proven wrong.

For instance, when that fix for the mangled data when up, I'm positive the majority of people who needed to use it never knew it was there. It wasn't mentioned in the LT blog, and the LT staff didn't post about it in any of the specialized language groups. Individual users had to bring it up on LiveJournal and in the LT groups; it was passed around strictly by word of mouth, and most people weren't lucky enough to be reading the right places, at the right times. I've also posted in a number of places about working with the ISBN thing, but people are still popping up constantly who haven't figured out why their covers aren't displaying.

That's certainly a pity. In general, I think it's safe to say that LibraryThing is at the point where it desperately needs some in-depth user documentation; it's not really sufficient to rely on one-on-one, word-of-mouth, anecdotal instruction any more. A Wiki would even suffice, if Tim & Co. don't want to take on the documentation task themselves. I'm a technical writer by trade, and online help is my bread and butter, so I get absolutely twitchy at times when I think how overwhelming it must be for new users who join LT these days.

Ah well, that's enough plaintiveness for now. At least it does sound as though the Amazon Japan ISBN issue may be tackled soon. If it is, I will be largely mollified. And if it's not officially announced anywhere public, let's make a pact - the first person who notices that it's fixed must share the news here!

Well, in that case you should probably see Tim's announcement here about LT in different languages which I think implies dealing with the character set issues.

Sorry, dropped a digit off the end of the link, now fixed

There's a number of us technical writers and editors here. I think that if and when they open up a wiki, there will be a lot of people willing to help develop user documentation.

Well, in that case you should probably see Tim's announcement here about LT in different languages which I think implies dealing with the character set issues.

That link points to the Jack Kerouac forum, and I don't see anything from Tim on there...?

let's make a pact - the first person who notices that it's fixed must share the news here!

Got it.

In the meantime, here is another new thread to watch: Swedish character display problems.

As for beating drums, any recommendation of LT from me would have to be an honest one, with the caveats "Keep in mind that some unicode stuff is pretty screwed up" and "Use at your own risk, and please don't blame me later, OK?" That wouldn't really be a recommendation per se. So I don't. If people I know get here on their own, I'll give them a hand in dealing with the inevitable.

I do find this sad because I have a selfish desire to see everyone I know dump their libraries online so that I can page through them and look at what they've got. I'm nosy that way.

Tim's announcement here about LT in different languages which I think implies dealing with the character set issues.

Um. An interface translated into Chinese doesn't mean that the underlying site is actually supporting Chinese. Y'know? And translating the interface into languages that aren't being fully supported by the site itself is . . . not very nice. New users will get the wrong impression; and the existing pile of frustrated people will grow ever larger; and the more, the merrier is not the case with this. If LT isn't going to fully support these languages, it needs to state that up front rather than implying the opposite; we'd stop pleading for tech support and repairs we'll never get. It's the cloud of unknowing here that's been driving everyone up the wall.

Those characters set issues are extensive, incidentally. Author and title links, none. Searches, none. Sorting, none. Mismatches and trashed information pages, rampant. The lack of ISBN transfers from Amazon Japan has been that way since that link first went live, so this segment of the site is already damaged in ways that we cannot tackle on our user ownsome. When many earlier users couldn't get their covers to appear they didn't investigate why -- they just walked away. That means a lot of trashed information pages are never going to be fixed since they were created by the first person to enter that work. So everyone else with those works is stuck with the trashed pages -- not to mention everyone who's being erroneously combined with them. (And you can't separate mass erroneous combinations. I've tried. It locks down and makes the bad matches permanent.)

So, I guess we'll see what happens. It still says "BETA" up there, so good things could happen at some point far down the road. Glass half full!

It's coming!!! Internationalization!! Or at least I think so.

See http://www.librarything.com/talktopic.php?topic=2180

Tim has graciously mea culpa'd to the communication problem (with the Frankfurter Buchmesse a little over a week away, LT staff must be going crazy about now), but I'm left wondering...did he forget to "hit" the Amazon Japan ISBN problem as mentioned in this thread or was an attempt to fix the ISBN import unsuccessful?

Has anyone tried this recently to see if it's fixed?

Tried it just now by searching on Amazon Japan using only the ISBN (4529035654) as a search term. The book is successfully added to LibraryThing, but the ISBN is not carried across, nor is the cover.

I'm still flummoxed as to why LT can't retain these details once the new record is added. After all, a thumbnail of the cover appears in the search results list prior to adding the book, and the ISBN is the only thing I used to (successfully) retrieve the other book details in the first place.

Try it yourself... it's a book on kitsuke (how to put on kimono), ISBN:
4529035654

After examining the XML output from a "web services" request to Amazon Japan, I think I may have identified the problem: webservices.amazon.co.jp doesn't return an ISBN element under Item Attributes (.com, .ca, .uk, and .de do). Since LT is probably using the same code to process data from all the Amazons, this oversight presents a problem.

LT could create a separate process for amazon.co.jp searches and import the ASIN ("Amazon Standard Identification Number") which is the same as the ISBN for books on Amazon. However, the ISBN-13 transition is only two months away. Other Amazon sites transmit both the ISBN (ISBN-10) and the EAN (ISBN-13) under Item Attributes. I assume LT will soon add an ISBN-13 (or EAN) field to their database and import both fields. I doubt Amazon intends to update the ASINs for all their old inventory, so how would LT handle an ASIN which could be either ISBN-10 or ISBN-13? Calculate the length of the string during import?

Okay. I finally solved this one. Thanks, Mangaroo, for pointing out the problem. The second insight was that LT now has a good ISBN-checking algorithm, so, basically, it checks the ASIN to see if it's also an ISBN. If it is, it uses it. If not, it remains blank. The point is to avoid filling LibrarThing up with Amazon numbers.

Looks like my :CueCat and I have a very busy day ahead of us. Thanks, Tim!

ETA: Oh, dear. It appears to work when adding a single book, but not for the universal import (where it's really, really needed).

A couple of comments.

First of all, thank you to both Tim and Mangaroo for explaining the problem and potential solution.

Tim, thanks also for making it possible to add a single book with ISBN and thumbnail cover remaining. This is terrific news! I tried it just now by searching only on Amazon Japan using the ISBN/ASIN 4529035654, and it worked beautifully. Even we haven't got the universal import yet, this is still a BIG advance!

However, there does seem to be a new quirk with adding books via Amazon Japan. When I searched on this book (4529035654), LT skipped the usual intermediate step of showing me the "1 search results" (to confirm whether it was the one I wanted), added the book to my catalogue, and showed the title at the top of the Recently Added list!

I wondered whether this was a hiccup that had appeared for all situations where a single search result is found; but no, when I searched for an English-language book on Amazon US, I still saw the (single) search result in the search results list.

So this is an interesting twist. I tried adding another couple of Japanese books via Amazon Japan - both of them have unique ISBNs/ASINs - and the same thing happened. In other words, LT immediately displayed the Recently Added list, with the searched-for book now at the top, rather than showing it on the Results (from amazon.co.jp) list first.

I don't know if this is unique to Amazon Japan... but it's an interesting wrinkle I thought you'd want to know about.

See here.

I think people may convince me to make it an option. I'm wary, however.

Ah, so it's a feature ;-)

OK, will comment in that thread instead of here.

In the meantime, though... thank you for finding a solution (on the single-book level, at least) to the Amazon Japan ISBN/cover problem! I can't express how delighted I am. {dances with joy}

Ohhhh, this is fabulous. I'm so thrilled. Good things happen when my computer's down for a week. Yes! (But I'm not going to trash my computer again to test that idea.)

Happy happy. Thanks everyone. Happy happy.