Dictionary for Sony PRS500/5 (autorun from SD card)

[spooky69]

I've been away from internet access for a while but I just thought I'd drop in and say I'm glad to see that this project had been getting a lot more attention. I won't be able to try out the new versions quite yet but I'm excited to see that a lot of the issues have been resolved. I haven't read through the newer posts extremely carefully, but in case it hasn't been addressed yet I'll just say again that I'd really like to see the a different control scheme for the 505 where the page turn buttons (of which there are two pairs) are used to move the cursor on the keyboard and not for the "next/previous word" function.
Besides that, I'd just like to thank everyone who's put work into creating a usable dictionary for the Reader. It's definitely something that adds to the usefulness of the device and you guys have come up with some really solid improvements to the early versions. I'll try to check back in soon with some feedback about the new version(s).

[joeanne12]

found out what I did wrong and thought I'd share my stupidity incase anyone else did the same thing. I have learnt not to do these things at 2am when you are getting tired!!! I put the wrong files onto the wrong card!! Had to do a hard reboot to get into my reader again. Have since put the correct files in the correct place and all is working fine now. Although still had to do a few soft reboots before the reader liked the dictionary and didn't freeze?!? So thank you very much for a great program.

[spooky69]

Hey, I just tried out lisak's newest version (after posting a report about the old version, which doesn't work at all on the 505 because the enter key does nothing).

I really like the new version (which needs to be linked to in the original post instead of lisak's first one) and there are only a couple of things I'd change. There's a rectangular "null" character at the end of each definition. I said this on the first page but I noticed that the default dictionary file when I installed the new version was now wb1.txt and wondered if there might be an updated dicitionary file that fixes this. If that exists, I'd suggest repacking a new archive with all of the files included and editing the first post to refer to that.

I also think it's essential to have some way to browse through longer definitions, since they seem to end up rolling off the screen. The up and down buttons on the directional pad seem like they'd be perfect for this, though I imagine it might be a little tricky to program.

This is just a small thing, but I might suggest having the first letter input after looking up a term start a new line automatically instead of having to clear it manually (though still leaving the definition on the screen), though that depends on whether you think it's more likely that the average user would want to start a new search term (like they might do when they're trying to find what cuss words Webster left in) or whether they would want to add more text onto what they already searched for (like they might do if they accidentally hit the search button before finishing their "typing"). I think it would be a little more handy if the line cleared automatically as you started inputting more letters, but it really is a very small detail, especially with the bookmark button serving to clear with line with one button press.

Anyway, everything is pretty much where I need it to be to be usable and here's a big THANK YOU for changing the button configuration around. I also like the cursor at the top and bottom of the screen (I forget if this was in any older versions) and non-qwerty keyboard layout. It's fast enough now that I don't it needs an hourglass, and the only reason it needed one in the first place was because it was too easy to accidentally hit the enter button and have it seem "frozen." If you did find a way to put one in, go for it.

Aaanyyyyway, keep up the good work, everybody, and thanks again for making my Reader even better!

[lisak]

Thanks for the report & suggestions! Just a few quick comments from my side:

There's a rectangular "null" character at the end of each definition. I said this on the first page but I noticed that the default dictionary file when I installed the new version was now wb1.txt and wondered if there might be an updated dicitionary file that fixes this.

Do you use dictionary file from http://dictiolaunch.googlecode.com/files/wb.zip and
version 0.85 (I guess so, but I'd rather ask anyway) - I don't see the character you're talking about on my Reader now, though I know what you mean because I remember this from one of the older versions.

(the "wb1.txt" is actualy just a remnant from some testing, it has no real meaning whatsoever...)

It's essential to have some way to browse through longer definitions, since they seem to end up rolling off the screen. The up and down buttons on the directional pad seem like they'd be perfect for this, though I imagine it
might be a little tricky to program.

This is an excellent idea and it should not be too difficult to get it there. I originaly thought about changing the font for the longer definitions, but this is better solution (and probably easier). Thanks!

...having the first letter input after looking up a term start a new line automatically instead of having to clear it manually

I understand, though it would be probably wiser to have this included in some '"user settings" dialog I'd like to create. The problem is that we're slowly running out of usable buttons - and I still don't know how to use Zoom and Volume keys. <– never mind that, I found it

[spooky69]

That all sounds good to me. To answer your question (and my own, now that I think about it) I was using the 0.85 version of the program with the original dictionary text file I posted on here a few pages back (I deleted the whole directory when re-installing the 0.85 version and I only had the old version at hand). I suspect that if I went back and used your wb.txt file again, the problem would go away. [Just tried, and it did fix the problem]

A couple of other hang-ups/suggestions/musings:
-When I pop in the SD card I need to hit a button (any button, seemingly) before the program will actually pop up. Otherwise, it will just sit on the same screen. I don't remember if the first version of your re-write did this, but I know that Clemenseken's didn't. It does still load automatically when I start up the Reader, which is actually another little something that I'd like to see gone, though I understand why it happens.

-Yikes, just noticed this one: I also have to hit a button (any button again) after searching to get the definition to pop up. This was actually going to be another error to report, because it seemed like some of the results were showing up much slower than they should have (like "ant" and such). For now, I can just hit enter twice in a row and it will pop up after it finishes the search. On the upside, I can see now that this search is actually much, much faster than I thought it was when I posted my first impression.

-Speaking of start-ups, is there any way to get the actual script that's running onto the Reader itself? You may remember a post where I found that leaving the program open drains the battery very fast (about halfway down in 2-3 hours on my 505, I think it was), probably because the program is cycling while it waits for input. I haven't tested for any battery drain on the newer versions, but if it's possible (and doesn't impair the start-up time), it seems like it might be a good idea to move all of the program's functioning to the internal memory, with the autorun script on the card serving only to point to the script that's on the Reader and launch it.

-Finally, is there a way to make the "exit" button not also send that input to the regular Reader firmware? In all of the versions, hitting "menu" to exit program has also made the corresponding screen pop up when I'm back in the "main area". Since this is almost always going to be while I'm in a book, it has the potential to cause problems. Last night, it took me back out to the table of contents I was using on my reader, and sometimes you can't get back from that screen to where you were in the book without having to re-navigate back to that spot.

-Okay, wow, here's a whole big 'nother thing: several of the problems (both of the "have to hit any button to get it to work" things and also the "menu button" one directly above) seem completely fixed when I start the program over again. I am working on trying to reproduce these again, but it does seem to have to do with it running differently after the first time...okay, so...

There are a lot of variables here since I've been re-installing and changing dictionary files around, and I can't reproduce these errors again yet (restarting and doing a device shutdown didn't do it; reinstalling might but I'm about done messing around for today), but I think the fix has something to do with doing my first search after installing the new dictionary file and then restarting the program, since the errors happened consistently until I did that, at which point they stoppd. If this becomes a problem, I can be a little more thorough in trying to track down the caustation of the problems I had since I pretty much remember all of the steps that led up to it. Thing is, I'm pretty sure all of the errors that disappeared were happening last night when I was using 0.85 and the old dictionary file...I'll maybe report back later after doing a full deletion and re-install with 0.85 and the current wb.txt.

So, how's that for a really long, confusing list of interconnected troubles? I did just realize one more that's very easy to fix: hitting the "left" button on the directional pad when you don't have a search term (like I keep doing since I was used to old button configuration [the new one works much better for me by the way; thanks again]) makes it "freeze" up for about 10-15 seconds as it looks for a previous term behind nothing. You should just disable the "previous word" function until they've done a search (and it doesn't do the same "freeze" if you try to go to the previous word from the first line, so that's okay). You might still want to leave in the ability to use the "next word" button without having done a search, though, since that has some slight usefulness if somebody had a short list of somethingerother and wanted to scroll through it. In fact, you might just want to make the "previous word" button go to the last line of text file depending on how long the delay is...in fact, I'm thinking that this whole situation would be best fixed by disabling both the next/previous word buttons until after a search, since if somebody's pressing those right off the bat, it's going to be a mistake 99% of the time and any delay while that line is dregged up out of the text file is going to be frustrating for the user who isn't trying to look up "a" or "zythum". Was that a long enough explanation of my thoughts on the next/previous buttons and their role in the dictionary program before a search has been done? Maybe this is a problem best left to a future "user settings" feature.

And one other thing: after using the dictionary a few times, I really think that automatically clearing the input line when a new letter is entered after a search is probably the best default option. Leave up the definition in case the new input was an accident, but I've already had several instances where I spend a few moments typing out a new term and look over and find I'm about ready to search for dogrock or saxaphoneniggard.

So, there's a bunch of stuff. I'll try to pin down what was causing those problems that went away, but that will have to wait until I'm less frazzled from typing about this program so much!

Oh, and of course I just had to think of one more thing: I already said that the new version needs to be the one that's linked to in the first post (old one doesn't work at all on the 505, and I didn't check to see if Clemenseken changed this yet so forgive me if that already happened), but I'd also suggest that the next version's archive include a text file with installation instructions and a short user manual, since otherwise people might have to track that information down in the thread. OUT

[lisak]

Ooooh, heavy report! - my comments are below

-When I pop in the SD card I need to hit a button (any button, seemingly) before the program will actually pop up. Otherwise, it will just sit on the same screen.

I know - this always happens for a while, then suddenly the problem mysteriously disappears and program is starting normaly. Somehow I'm in doubt whether the code itself has to do anything with it - actualy, the dictionary screen should pop out before a program code itself is executed. I'll try to do some tests to find out where exactly it hangs up.

I also have to hit a button (any button again) after searching to get the definition to pop up.

Hmm, I haven't encountered this one yet.

it seems like it might be a good idea to move all of the program's functioning to the internal memory, with the autorun script on the card serving only to point to the script that's on the Reader and launch it.

I know about the battery drain (I learned it the hard way - today I left my reader open with Dictionary running during the night and in the morning the unit was completely dead w/ some weird residuums at the screen - it scared the hell out of myself), but I am not sure if the sole fact that application is running from the card is causing this.

Anyway, I can shift autorun.js to internal memory, but autorun.xml has to stay where it is (I can use symbolic link to redirect it elsewhere, but card would be accessed anyway). I am still trying to find out how to start application without using autorun functionality, but it's only realizable through a hack, which is rather risky with 505 and somewhat difficult.

Finally, is there a way to make the "exit" button not also send that input to the regular Reader firmware?

I am afraid that the answer is "no", because I believe that "exit" from the book actualy happens even before dictionary is launched (when you insert the card, current viewer is interrupted, "main" application pops into foreground and then content of autorun.xml is evaluated). And I suppose that if we ever get the dictionary running without autorun.*, you won't be able to start the dictionary while reading at all.

...several of the problems (both of the "have to hit any button to get it to work" things and also the "menu button" one directly above) seem completely fixed when I start the program over again. I am working on trying to reproduce these again, but it does seem to have to do with it running differently after the first time...okay, so...

See? Told you so. Don't be afraid, these glitches will return. But if you do find out what combination makes this one pop out, let me know.

in fact, I'm thinking that this whole situation would be best fixed by disabling both the next/previous word buttons until after a search,

Good idea. I still consider saving the last searched term (maybe last nine of them and access them via buttons...?), but I am not sure if anybody would find it ever useful (think 80/20 rule). This idea of yours might be a good quickfix.

Oh, and of course I just had to think of one more thing: I already said that the new version needs to be the one that's linked to in the first post (old one doesn't work at all on the 505, and I didn't check to see if Clemenseken changed this yet so forgive me if that already happened)

This is up to Clemenseken... I'd recommend to put a link to http://code.google.com/p/dictiolaunch/ there, where the latest version resides (alas, for 505 only...) and where also documentation will should be stored. :

[=X=]

Originally Posted by spooky69

-Finally, is there a way to make the "exit" button ...night, it took me back out to the table of contents I was using on my reader...

Couple questions here. Are you double pumping? Hitting the button more than once.

I know one annoying artifact of using the Dictionary is that it must reside on the SD Card. When I am reading books that are located on the SD Card I tend to lose my location. It seems the Reader only saves the "current location" when you exit your book. One trick I've learned to get around this is to bookmark your location before calling the dictionary. At that point is will be very easy to get back to your location.

Spooky, while having the application auto clear the input box is a great feature for testing the dictionary, in practices it is more of a drawback. I find that when I'm actually using the dictionary to lookup the word I don't want it to clear the input box. Most words I lookup are very long by the time the dictionary loads I've already forgot how to spell the word. So I typically enter half then enter a few letters at a time to get closer.
We, however have two methods to enable a user to quickly clear the input either through the keyboard of from a button.


@lisak
Whith the application getting I wounder if it would be worth while to improve the installation process. I'd recommend we put ALL the files in one directory. and instruct all users to unzip the Dictionary-vX.zip into the "Sony Reader"
The Dictionary zip file would have a directory structure of so.

autorun.xml
autorun.js

+Dictionary\Skins

+Dictionary\<ALL Dictionary Files>

The autorun.xml file would have to be modified to point to the "Dictionary\Skins"


Any thoughts?

[spooky69]

Originally Posted by lisak

See? Told you so. Don't be afraid, these glitches will return. But if you do find out what combination makes this one pop out, let me know.

I thought I'd found one (turning the page in a book), but it doesn't seem repeatable. I did notice a trend of, after turning a page, having to hit a button to get the program to pop up, and then having it come up fine when I do it again from the same page in the book. I'm still not sure that's not a real phenomenon, buuut...

After trying this a few times, it seems like it's almost 50/50 as to whether any individual insertion of the card will make the dictionary pop up correctly. I am pretty certain that the error was happening consistently for a period and then seemed to go away almost completely. Having just tried again from the main screen (without changing the screen, and having had it work the first time), the error popped up about two out of 10 times I put in the card.

If I had to guess, I think it has something to do with the computation/input cycles of the Reader itself. I can't determine any pattern right now, but since the error happens without warning during a series of identical "tests," the conclusion would seem to be that popping the card in at certain moments causes the program to launch differently (or not launch at all, as I've noticed happening with all of the versions). So, it seems it's possible to catch the Reader at a bad moment in its processing cycle, which is the only possible determining factor I can pin down.

I don't remember this "have to hit a button" thing happening with Clemenseken's old versions (of which I tried two, I think), so it might be good to go back and compare the "launching" aspect of the code and see if there are any changes there to explain what happened. Keep in mind that I can't guarantee that there wasn't this exact problem Clemenseken's version and I just didn't notice it or notice it or encounter it.

Another error that I got during my testing, which I haven't been able to repeat at all, was when I accidentally exited the dictionary by popping the card out, it took me back to first page of the book. Moving to a later page and exiting the dictionary by either hitting the exit button or popping the card out would keep taking me back to page one. It kept doing this and didn't stop until I went and tried it on another book, at which point the problem went away and didn't come back when I re-entered the same book it originally popped up in. I'm not sure that this was caused by popping my SD card out, but it does seem to have, once again, shown that the launching of the dictionary can interfere with normal processing cycle of the Reader (in this case somehow setting the current page to "1" and resetting that value/not letting it change until I exited the book). This book wasn't on the card, but it sounds similar to the problem =X= described (and I wasn't double pumping the exit button, although that problem has gone away completely and I'm not sure that I wasn't just describing a phenomenon that only existed in lisak's first version, or was being caused by using the mixed-up dictionary files...I'll have to do that full re-install soon and report back on any changes). Until then, other than the oddball error I just described, exiting the program has taken me right back to the page I was on with no trouble.

It seems to me like the most likely route to a solution to this problem is to somehow insulate the program from the Reader firmware's functioning and/or the card itself. You said that you could move the autorun.js file onto the reader by having the xml file point to the script on the internal memory, which seems like a likely solution to the battery problem and might just fix a lot of the other errors I've been describing. I'm pretty certain that it is the script running off of the card that's causing the battery drain, since accessing data off an SD card is a known drain on the 505's battery (people on here were reporting that they only get a few hundred page turns per charge when reading books off of an SD card, with the same problem true to a lesser extent when reading off of the Duo slot). As you can imagine, any active accessing of the data on the card (like it would do as it runs the script, unless it's running from an internal buffer or something fun like that) is going to cause that same kind of drain, so having the actual script moved to internal memory should give everybody using the dictionary a little more battery life at the very least.

As for the line clearing thing, that explanation does make sense to me, and I'd say that you're probably right that in thinking that leaving the function the way it is going to be essential for a lot of people. I do think it's something that should go in any future "settings" menu.

Just one other little thing that can wait until later: the "> PAGE < <Cursor> next/previous word" thing at the bottom needs to be fixed, or most likely should just be taken out to make more space for the definition. I'd also argue for taking out about half of the grey blank space up toward the top for the same reason, though all of this can obviously wait until a more finalized version.

[lisak]

we put ALL the files in one directory. and instruct all users to unzip the Dictionary-vX.zip into the "Sony Reader"

Well, at the moment I have to have dictionary itself in the internal memory of the Reader - it's the only way I can access it from the internal shell so far. And if it's really a fact that reading from the card drains battery, it would be probably wiser to let it stay there. Some installer would be probably possible, though (eg. copy everything onto card and move dictionaries to the internal memory at the first start) but I am still wondering whether it's worth the effort, especially when we're able to help people with installation problems here in forum.

It seems to me like the most likely route to a solution to this problem is to somehow insulate the program from the Reader firmware's functioning and/or the card itself.

I am afraid that it might be almost impossible thing to do, as the program is running basicaly as a part of PRS internal "firmware" application. I guess that the glitches and problems we're encountering are happenig simply because we don't know how to pop-up application "properly" (how to set all appropriate variables etc.), but honestly - to debug anything on the unit is quite PIA and debug something that's hard even to reproduce consistently is pretty frustrating - eg. today I spent good 30 minutes over a single trivial typo in the code that would be spotted immediately under normal circumstances...

Anyway, enough whining, on with new version:

DictioLaunchX-0.87:

* Added scrolling for long definitions (up/down arrow)
* Left/Right arrows disabled before first search (source of lags)
* If you feel like that, you can change clearAfterSearch from "false" to "true" in autorun.js - that way the input line will be cleared for you after you enter a letter after you finish search. Somebody likes it, somebody doesnt - for the time being, default is to have this functionality disabled (you still have Mark button for quick clearing of input line), in some future version it might become settable via user settings.

[=X=]

Originally Posted by lisak

Well, at the moment I have to have dictionary itself in the internal memory of the Reader - it's the only way I can access it from the internal shell so far. And if it's really a fact that reading from the card drains battery, it would be probably wiser to let it stay there. Some installer would be probably possible, though (eg. copy everything onto card and move dictionaries to the internal memory at the first start) but I am still wondering whether it's worth the effort, especially when we're able to help people with installation problems here in forum.

Ops the complete thought was in my head I just forgot to write it all down.

My thought is during the loading of the program, all files that must reside on the internal memory are "Installed" (e.g. copied) during to the internal memory. This would make this process invisible to the user. They would just notice a large load time for the very first time they run the program.

We can minimize the load time by performing a "file exist" (getFileInfo) check for each file. Or write a flag to a file when the a successful install was completed therefore we write a true flag. Then check for the flag.

[spooky69]

Very cool. Thanks for implementing the scrolling feature and also line clearing! I do have one more little request for the next version which would work well as a editable variable in the source code: making the scrolling take the form of starting the new block of text where the old one left of, i.e. scrolling by page instead of line. That's how I pictured the scrolling taking place, although the way you did it seems to work very well and might be the way I'd prefer it if I could try both. If it's not too hard to implement in the next version, I'm sure it would be a good option to have both available.

I have noticed one really, really small thing, which is that some one-letter searches seem to take a long time, especially searching for "A," which takes about 20-30 seconds. I'm sure this is probably a function of the grep search and can't be fixed, not that it needs to be, but I guess I'm just trying to keep an eye out for aspects of the search that might represent some lack of optimization. Also, the Webster's dictionary file has a few words that are out of place alphabetically and should be fixed since they can mess up searches. There's surely a script that can do that automatically, so I'll look for one sometime soon and post a modified dictionary file when I do.

Anyway, I don't want to sound down on the search at all since it's been massively improved and has let me start looking things up in texts I'm reading...I now know what the words "ermine" and "ducal" mean. Sooo thanks again so much to Clemenseken, lisak, and =X= for putting this thing to together! It's a really great addition to the features on the Reader and will help me out a lot while reading the show-offy vocabulary found in "The Works of Edgar Allan Poe" from the Sony Connect store (best $0 I ever spent, although it's only $2 in the classics section and you get quite a bit for your money). Also, check out Aesop's Fables, because they are cool and you will learn how not to get eaten by various animals.

[naeem]

guys, how do i send files to the prs 500 internal memory. Is it meant to come up as a removable device on the windows system.

[pilotbob]

Originally Posted by naeem

guys, how do i send files to the prs 500 internal memory. Is it meant to come up as a removable device on the windows system.

The 500 does not mount as a storage device. You need to use the eBook Library software... or alternately calibre.

BOb

[spooky69]

Originally Posted by pilotbob

The 500 does not mount as a storage device. You need to use the eBook Library software... or alternately calibre.

BOb

Busted.
To contribute, DictioLaunchX makes reading Shakespeare on the Reader about .7% better, that being the fraction of the words in the thing that I wouldn't be able to understand without looking them up! Should there be any serious concern in my brain about breaking the SD card slot's spring (I'm using my rudimentary knowledge of things in things to guess that there is some sort of spring thing in my silver shot arkansas), because I've seriously already popped my SD card out ~300 times, and back in ~301? I seriously feel like I'm having sex with my Reader really slowly, which isn't a sexual turn-on, but, once again, is actually a real concern about whether I might break the damn thing. Maybe we should start worrying about making a hack-centric menu-based thing or some shit?
Either way, I can't help but feel this project is about ready to creep out of the shadows of the Dev Corner and into the mainstream superforum of the regular 505 board. It needs to happen once the program reaches a "throughly usable" status (if only for the usefulness to the crazy old people on here who are like, "Computer codes?!? Noooooooo!", meaning they would never come in here and be able to make use of the program), and I'd suggest that lisak post the new topic if you have the time and patience and inclination to deal with a bunch of people thanking you a bunch and being really appreciative also demanding that you spend your leisure time programming a program for their luxury gadget which lets them read books from the internet.

[spooky69]

Really?