I thought I'd post my requests here but then again I guess other people might have some too so I guess you're all free to post whatever request (for DJ) you have.

Ok here's mine:

* official online page for kanji-crippled phones
* regexp-ish search ability. I guess there is a bit of it in there already, but it's not documented and if I try something funny it just bombs out. I think this wouldn't affect usability because it would really be invisible, and just there if you read the documentation well enough. :-)

That's all for now...

Posted By: Jonny

* official online page for kanji-crippled phones

That is probably not so hard to do, and something I'd like to have. However I'd prefer it if everyone just used iPhones :)

Posted By: Jonny

* regexp-ish search ability. I guess there is a bit of it in there already, but it's not documented and if I try something funny it just bombs out. I think this wouldn't affect usability because it would really be invisible, and just there if you read the documentation well enough. :-)

This I'm not sure how useful it would be as the current wild cards cover most search cases. Can you provide a specific example of something you would like to be able to search for with regexps? The problem with them is that they can be very slow, and it would be a lot of work for a potentially not very used feature.

Posted By: tacomanator

This may be out of the question, but I would sure love to be able to look up meanings in Japanese. Is there an Edict equivalent for a 国語辞典?

I would like that as well, but as far as I know there is no such dictionary data freely available.

The wildcards are documentet, but I am aware that it is at a less than obvious place. If you go to the word search by clicking the menu, that is without searching, you will have a small documentation text below the search area. But since most users come to the word search by doing a search on the front page they never see this text. I am aware of the problem and intend to do something about it.

For completeness, here is the small blurb on wildcards: Use the wildcards * (zero or more characters) and ? (exactly one character).

I just remembered something.

A way to search by codes.
Example: Like you could enter the code for adjective, then enter a kanji, and then get all the adjectives using this kanji.

Jonny: Yes, that is a good idea. It was actually possible way back in an earlier version, but the code was ugly and incomplete. I intend to add that feature again when I finally get around to using JMdict instead of EDICT.

YuanHao: Definitely a good idea! I have been ridiculously busy with life since I launched the forum so I haven't been able to give the forum the love is deserves. I'll try to get to it as soon as possible.

It makes sense if you want to be able to make very specific searches, like search for all English words containing "rice" where the kanji contains "米" or something similar. This is what I inteded it for in the beginning, but I agree that this is not the way it is used in most cases. So changing the default behaviour is probably a good idea. However, I would rather consolidate the two boxes into one, and intelligently parse it to see if the user is searching for only one language or both English and Japanese, but this will screw up automatic romaji detection in some instances, and I haven't toyed anything with it in order to come up with a good compromise yet, but it's on my todo-list.

Posted By: Brutulf

Hi there. I just stumbled upon this site, a great resource!
That said, I just noticed that the search box is actually case-sensitive...which meant I was unable to find many words when I just started out, before I realized this. So to my point, an option for making the search case-insensitive would be nice, and especially helpful when copying and pasting from elsewhere. :)
Keep up the great work!

Japanese input engines are always case-sensitive. It makes no sense to make it case-insensitive.

Sorry, I should have specified. I only know latin characters, so I referred to romaji search on japanese words. I guess what I'm wanting for is an "ignore uppercase" -checkbox, if that's possible. :-)

As mentioned, it's handy when copying/pasting from other sources, where not everything is in lowercase. No big deals, just makes it potentially a wee bit quicker, as you don't have to "fix" the first letter if it's capitalized.

What I would love to see on jisho, is the verb conjugation charts that are avaliable on the wwwjdic site. It's the only reason I ever visit wwwjdic nowadays.

I'd like to see a software with these features:

* probably using Java
* to be used in mobile devices (phones)
* has an offline copy of edict (or any other dictionary, but edict is a good start)
* ability to update the above dictionaries through a one click button
* has a killer interface ;-)

The offline dictionary feature would be good because Docomo charges per bit (and I guess many other phone companies do as well) and a dictionary easily fits into the memory cards of our days. Also when on a train without 電波 the dictionary would still be usable.

Posted By: dpihl

Hey Gang! Hajimemashite.

Sorry to bug you all, but I'm at my wits' end.

Many of my Japanese American friends are about as Kanji illiterate as I am, and many don't even speak Nihongo.

Yet genealogy continues to grow as the most popular hobby around.

I volunteer at a Family History Center, and was helping a friend trace his roots when I stumbled onto the problem.

His American ancestors were easy to find. But each time we crossed the pond (so to speak), the trail went cold.

Not hard to figure out why. There may be plenty of Kenji Nakamuras here in the US. but there is no such person in Japan.

Are you kidding me? A Romanized Japanese name in a phone book, death certificate, grave yard, etc??? Not very likely, and I'm not just talking about Hepburn Romaji vs. some other flavor.

If Denshi Jiten can be integrated into Firefox or Safari, then what about the search window of sites like Gencircles, World Vital Records, or Ancestry.com?

I'm no programmer, so I have no idea what this would take.

Seems to me the ideal way to do this would be to create the appropriate name associations. Sort of like Soundex, Metophone, and NameX.

You search for somebody with the surname Ito. You get results back with all the romanized variations. Itoh, Itoo, Itou, etc. You also get back any relevant hits in Kanji.

Any ideas how a non-Japanese speaker could search for his Japanese roots online?

Basically you want a search engine that searches all romaji standards for a certain word? So if you input Internet Hepburn (Itou) you would get Standard Hepburn (Itō)?

While I certainly see the usefulness from an academic perspective, I doubt there's incentive enough to make a site aimed at this goal. First of all, there's only two big romaji standards (hepburn and nihon-shiki) so there's usually no reason to search since it's easy to know both. Second, most people on a decent level of japanese (the people who would use the site mainly) very rarely use romaji.

Posted By: Richard

first of all, Denshi Jisho doesn't have names in at all

Remember, Denshi Jisho has an included Names dictionary, which does just this (Find all the different variations in kanji of a name).

Thanks for all the suggestions and your requests for clarification.

A simple web page containing all Kanji and Romaji variants for a single name in itself is bound to be helpful.

After all, it is a starting place. You can then try to memorize the Kanji associated with your surname (assuming Japanese ancestry).

You always start by actively searching the major Genealogy web sites. Thanks for the extra links Richard.

Simple cut and paste routines would allow you to populate the "advanced search" windows over at Ancestry.com, Familysearch.org and so forth.

Now that facebook supports Japanese, it probably makes sense to post a Kanji/Kana version of your full name on the facebook app called "We're Related". You never know who you might hear from.

As for the very thoughtful question about what one does with the results page if they are all rendered in Japanese, I have a few thoughts.

You might find a fully assembled pedigree going back to emperor Jimmu. Not very likely, but one can always hope. Even the smallish pedigrees you find would be worth looking at. With enamdict you can Romanize the chart one name at a time.

Secondly, there are a few genealogy utilities that build pedigrees using automation. GenSmarts is a fer instance.

If you can provide such an engine with a few basic clues, there may be a whole bunch of useful data in the results window.

Searching for Kenji Nakamura may be of limited use, just as searching for John Smith gives you waaay too many potential hits.

But if you tell the search engine that John Smith is married to Hepsubah Hoffenpoffer, every result you see will probably be relevant.

By the same token, there may be plenty of Kenji Nakamuras out there, but significantly fewer of them who are married to Yumiko Satoh.

If you enter Yumiko's details and also the fact that Kenji's father's name is Tomo, you narrow your results even further. Add a birth year to the equation if you have one. Add marriage details or anything else you have at your finger tips.

Eventually you'll narrow your results down to one or two possible Kanji Nakamura/ Yumiko Satoh combinations.

Post the details you dig up on GenCircles, OneGreatFamilyTree, Ancestry.com, or other collaborative genealogy sites. You never know who you might come along and graft entire branches into your family tree.

One might well ask where one finds marriage and birth information for somebody if one cannot read Japanese.

Whichever ancestor emigrated is bound to have left vital records with agencies on both sides of the ocean. Birth records would be held in Japan presumably. Death records are probably had somewhere in the country where the person died.

An ideal scenario might be finding a census record. Kenji would be listed as the head of household, and Yumiko would also be listed in the same household. Both ages would be recorded, and the country of their birth. Sometimes a parent's name is listed, or a mother's surname, or other details about previous generations. Often the number of years married is listed.

Sometimes you can find the ships manifest, which contains many of these same bits of information.

The more you know about the American side of the family, the more likely it is for a machine to automatically calculate the rest.

http://www.csse.monash.edu.au/~jwb/cgi-bin/wwwjdic.cgi?1C

Just noticed the wwwjdic has added enamdict to the list of available dictionaries.
This solves my problem.
Now if only I could remember which results are surnames, which results are male forenames, etc.?