Search function

[gandq]

Dear Mike,

i was trying to figure out the search option on the very right of the screen on top of the list (i am not entirely shure what it is supposed to to) when i had an idea i need to suggest:

a search option to show every entry that contains a certain charackter,
maybe with with an option to chose its position (any, first, second, third, ...)

this feature is actually implemented in Clavis Sinica, now that i think about it. anyway, it is one that would greatly increase the dictionary's worth for studying. i know that i can search with wildcards, but that is not exactly the same thing. by making it possible to search all entries that contain a certain character, plecodict would leave the boundaries of normal dictionaries behind and become the key to a broader understanding of chinese (i.e. 'clavis sinica')

i hope i could convey the idea. and i hope you will consider it (and i hope you will implement it).


keep up the good work,

cheers,




jo.

 

[mikelove]

Actually this has been discussed before, and will likely be implemented in 1.0 - it only requires some very minor modifications to the current wildcard search feature. The position-based search you mention is already possible with wildcards; simply enter the character itself for first position, a @ followed by the character for second, @@ and the character for third, and so on... (the wildcard character will be customizable in 1.0) It even works with Pinyin.

And this feature is not only in Clavis Sinica but in Wenlin as well; by the end of 2005 (PlecoDict 2.0) we're hoping to support just about every feature in either one of those programs. (and yes, this includes audio pronunciation, stroke order diagrams, and component-based character lookup, so you may begin salivating now)

 

[gandq]

JOY !!!



jo.

 

[goulniky]

Actually, the otherwise excellent Wenlin does not support wildcards : you can search for (Chinese) words starting with a particular character, or containing a character, but not specify at which position.

 

[mikelove]

Well of course we're happy to add some features that aren't in Wenlin too The new mixed character/Pinyin search feature (which is already working beautifully in our test builds) is another example of this, though you can achieve much of the same functionality in Wenlin by typing in Pinyin and using the "convert" button.

Wenlin is built around its text editor, while we're built around dictionary lookups - the search field is continuously available in our software, whereas in Wenlin it has to be brought up with a menu command. We probably will be adding at least a document reader in PlecoDict 2.0 (for faster definition pop-ups), but in general you can expect Wenlin to be better for viewing/editing text and our software to be better for looking up words. Which makes sense, since their software runs on desktop and laptop PC's and our software runs on portable devices.

But we don't really view Wenlin as a competitor, in fact we have a very good relationship with them - each of our products serves a different purpose, and there are numerous reasons why it would be worthwhile for someone to own both. Even if somewhere down the line we decided to develop a desktop version of PlecoDict, it would likely still be oriented around dictionary lookups and hence still be a very different product than Wenlin.

 

[gandq]

dear mike,

in the old oxford E-C dictionary there was a very good feature for searching english entries: if one wrote a word, that did not make up an entry by itself, but which appeared in the definition of the 'root word', the dictionary would find it there and highlight it. this seems to be not working in pleco dict. an option for searching any word in any location - entry or definition - would be very cool, especially for searching custom dictionaries (i ofthen input example senteces, so the first word is not the important word i would search for when wanting to access it). kind of like a wildcard search for E-C. i guess it would be quite slow, but it could be made optional or something.

cheers,



jo.

 

[mikelove]

Actually, we didn't bother implementing that feature in PlecoDict because between the small fonts and boldfaced subheadings we didn't think we needed it anymore (since it's easier to just skim through the relevant entry); it really wasn't a full-text search, it just searched the nearest few matching entries to what you entered and jumped to a match if it found one. If a lot of people complain about its absence in 1.0 we might add it back in in 1.1 or another update.

We have considered adding a full-text search feature, which would among other things allow you to search for an English word in a Chinese-English dictionary, but that won't be available for a while and I'm not sure if we'd be able to get it working efficiently with custom dictionary entries.

 

[gandq]

that's a pitty

i still think a full text search would make a lot of sense, especially for custom entries.

maybe it is possible to make the custom database searchable by the palm find application for a start?

greetings,




jo.

 

[mikelove]

Possible, yes, but I don't think it'll make it into 1.0.

 

[gandq]

resident module

hello mike,

today i discovered the great merrits of using the plecodict resident while reading chinese text files on my palm (i finally managed to import them correctly - took me some time). i have to say that it is a complete new way of doing some chinese reading and i guess i will use it a lot in future.

here is one little thing i noticed:
- the resident uses the chinese os to display the characters. i don't know about CJKOS, but with ChOS a typical entry looks like this: 1bi4ye4* 毕业 [毕业] ... - simplified and traditional characters cannot be displayed at the same time, so the same characters appear twice, unnecessary. nothing very important, i just wanted to mention it.

and a question:
- will the resident be supporting the flashcards, i.g. will it be possible to create flashcards from within a chinese text file by using the resident? (best would be two different methods: 1) creating a new entry from the marked text and 2) adding a dictionary entry from a resident search)


thanks for your time,




jo.

 

[mikelove]

We've already got an option in the works for the finished version to show only traditional or simplified characters (though you can also still show both), so this bug should be taken care of soon.

And we are planning to support flashcards in resident mode - we've managed to trim down the basic flashcard system code to the point where we can fit it into the resident-mode application, so it should be possible to create a flashcard from a dictionary entry that you're viewing in resident mode. Unfortunately, though, it will not be possible to add a custom dictionary entry in resident mode - the code for that is a lot more complicated and there's no way we'll be able to squeeze it in. (Palm OS limits a pop-up application's code to 64K, or about 1/4 the size of the full PlecoDict program, so we're very limited in the features we can enable - this is why we use CJKOS or ChOS instead of our own font engine)

We were originally planning to build our own text reader right into PlecoDict, but we ran out of time and weren't able to get that feature implemented in 1.0. It's still something we'd like to do in the future, though, and using a built-in document reader would give you full access to all of PlecoDict's features while reading a text file.

 

[gandq]

an implemented text reader that fully supports the flashcard feature is a great idea. you seem to be really eager to make plecodict the absolute killer app (at least for students of chinese). i really apreciate your work very much.

btw: how many people actually work at your company? one can easily get the impression that you do everything by yourself, posting news, administraging the forums, coding the software...

do you actually have time to continue studying chinese yourself?

i hope so, as your software makes it very rewarding.


many greetings,




jo.

 

[mikelove]

Well it's difficult to say how many, exactly... basically I run the business and do all the stuff which requires constant attention (like posting news, support, etc) and I also do a lot of the coding, but I have friends and contractors who help out at various times with development, file conversion, content production, etcetera. (part of the reason for our unreliable release timing is the fact that people sometimes get busy with their "real jobs") But if the next few launches go well, we may finally get to the point where there can be someone other than me working on Pleco full-time. I could probably farm out some of the support stuff to a third party, but it really isn't very much work and it's a good way to keep up with common problems / questions / etc.

And thanks for the compliment - I don't have much time for Chinese study these days, unfortunately, but I do practice with people occasionally to try to at least keep my basic abilities up.

 

[Jianchuanbrit]

Going back to the original post, it finally helped me understand what that little square meant with the numbers in it (1, 2 3 etc). I always thought you had to click on it to select the search position but now I see it is only there to show you what position you are searching in. So you have to put the at symbol in first. But could this search system not be by entering the search string then selecting what position you want to search for it in?

 

[mikelove]

We've heavily revised the list navigation system for the beta, actually - now there's just a single button which toggles between browsing the entire dictionary and looking at a list of search results. The dictionaries are now sorted in that combined pinyin/character method that most printed dictionaries use, so there's no more need for separate pinyin/character sorting, and if you want single-character mode there's an option in Preferences to have searches for a single character or syllable come up with single-character entries only. (though if people really complain about that change we might consider adding the single/multi toggle back in as a toolbar button)

And we've eliminated the position trigger too, so if you want to search for a character in the 2nd/3rd/4th position you simply enter one or more wildcard characters before it.

 

[gato]

Searching with multi-wildcard '$' followed by a characters currently results in a list of words starting with the character in the first position, then in the second position, then third, and so forth. What I really want is a list of words in which the character in the query is in the second position or later. A search starting with "$$" in the query string achieves that. But I think using only a single "$" should give that result, i.e. exclude those words starting with the character. In other words, "$" should mean "one or more wildcard characters", while "@" means "exactly one wildcard character."

Has anyone else seen this behavior?

 

[mikelove]

Actually, we deliberately made $ behave that way, because it provides you with more flexibility; if $ followed by a character only gave you words in which that character was in the second position or later, there would be no single search that would give you all words containing a character. Whereas with the current system, $ followed by a character will do exactly that, and @$ will still give you all the words in which the character is in the second position or later.

 

[gato]

Oh! Pretty tricky. Is it explained in the manual (not that people will spend the time to read about searches)?

 

[mikelove]

Indeed it is, on page 35 in the current version on Palm.