Search Function For Pocket Internet Explorer?

[kzemach]

Is there a plug in or upgrade (I've looked and can't find any) for Pocket IE that would allow one to search within a webpage for text, like you can in regular IE by doing a <ctr>F.

We could really use that functionality for searching for information on our units. The webpages are all off-line (on the actual iPaqs); perhaps there's some way we can embed an internal search function in the HTML???

I'm not optimistic, but if anyone knows of an upgrade or way to do this, it'd be great.

Thanks!

Ken

[Pony99CA]

Is there a plug in or upgrade (I've looked and can't find any) for Pocket IE that would allow one to search within a webpage for text, like you can in regular IE by doing a <ctr>F.

We could really use that functionality for searching for information on our units. The webpages are all off-line (on the actual iPaqs); perhaps there's some way we can embed an internal search function in the HTML???
If the files are in your Pocket PC's My Documents folder, just use Pocket Word. HTML files will show up in the file list, so open the file and use the Find/Replace... function in the Edit menu.

Steve

[kzemach]

Yeah, that will work. Sort of. The main reason we're using webpages is so that guys don't have to search through files and documents. The information we want to put in the webpages is currently an Excel file, so we could leave it there, but that means we need to teach them how to open Excel, and use the find function in it. It's a training issue, and a time to do those actions issue.

I know, I know, that sounds really easy. But we're talking soldiers here. We're talking Afghanistan. We're talking long lists of pictures of wanted persons. Lists of passport numbers of wanted persons. Lists of liscense plates of wanted vehicles. Phrases prerecorded in Pashtu so that they can communicate with the locals. It all works stellar, except for searching for the passport or plate numbers.

The HTML interface makes it REALLY easy for them. Not all soldiers know how to work a computer per se. Not all know how to open and search documents. But ALL soldiers know how to surf the net. Thus, we'd like to keep everything in the 'net form. Oh well. Maybe PPC 2003.

Hollister, huh? Gosh, makes me a little homesick (I'm from Redwood City). Kandahar's a hoot, but every once in while it'd be nice to go a day without dust up my nose.

[rhmorrison]

PIE understands Java Script language.

Wouldn't it be possible to do something in Java Script to perform the function that you want inside the HTML? 8)

[kzemach]

Really? Cool. I guess I'll have to learn Javascript now... If it works, that'd be stellar.

And a response from Germany now; flying through Frankfurt tomorrow! Very funny. The world is a small place.

[smittyofdhs]

ok kzemach I've read your post (very interesting hints you gave there) and I've checked out your site. Tell us...what do you do for a living? And how did you get involved in helping soliders with PPC?

[kzemach]

I'm a technical consultant; our company mostly does failure analysis of scientific and engineering systems in the states. We do very, very little military work (Ever tried contracting with the gov't? Better off slitting your wrists).

Anyway, I'm out in Kandahar on a program to assess technical needs of the ground troops and come up with solutions. We've done a lot of rapid technology deployments that aren't PPCs.

So, PPCs aren't our mission per se; I just saw that when searching a village, there was a serious problem with communicating complex instructions when an interpreter wasn't available. For instance, "We need you to move the women into one room, so that we can search the rest of the house." This isn't a very easy thing to mime, and if you know anything about the Muslim culture, the female issue is a very, very volitile one. Anything we can do to reduce tensions with the people we deal with is a great asset.

The original concept was to pre-record several (~40) phrases in Pashtu using the voice recorder, and then tap them later to say what you want. Worked like a charm. The soldiers love it, and we've moving to deploying 30 (one for each search leader in the Kandahar area) as we speak. But we'll need more; the medics use them for communicating with local patients, etc.

The interesting thing is, once you have a valid reason to have a PPC in the field (which is the translation), you can leverage that power for a host of other applications, like the pictures and the data. Looking at hooking up to GPSs, loading maps, and getting situational awareness. Keeping packing lists, equipment lists, firing tables, it goes on and on.

Ultimately, I don't think the iPaq per se is the ultimate answer, even in that weatherproof red case for a variety of reasons. But it is proving out that some form of PDA is enough of an asset that soldiers clammor for it prior to a mission; it's really helping them out. And the added bonus is that it's reducing tensions and problems with the locals. 99.8% of this country is filled with wonderful people; we've got to treat them with the respect and dignity that they diserve while still accomplishing the mission at hand.

As an aside, I knew NOTHING about PPCs prior to deploying these (I'm a PalmOS guy), but the users here at PPCT have given excellent guidance and help. This place has been great. Thanks folks!

[MikeUnwired]

The product is called the Phraselator. The software runs on Pocket PCs with ARM processors -- iPAQs were chosen because they had better speakers. The company also produces a rugged unit that runs CE.

http://www.phraselator.com

[kzemach]

The product is called the Phraselator. The software runs on Pocket PCs with ARM processors -- iPAQs were chosen because they had better speakers. The company also produces a rugged unit that runs CE.

http://www.phraselator.com


No, it's not. The Phraselator sucks/is irrelevant for what we need to do. The Phraselator is effectively software developed by DARPA, where the user speaks a phrase, the software recognizes what the user is saying (in theory, some of our guys in the States have tried it and beg to differ), then selects a phrase in the native tongue from a list of ~3,000 phrases that have been prerecorded.

There are effectively four problems with this approach.

1. The list of phrases was developed by someone sitting at a desk in the US. The soldiers here looked at the ~3,000 phrases that the Phraselator has, and liked TWO of them. All the rest were useless. Furthermore, the Phraselator has NONE of the most important phrases that the soldiers need.

2. To add phrases to the Phraselator is a time consuming process that involves developing a new phrase set plug-in on a desktop. With the way we're doing it, if a soldier wants a new phrase in the field, they grab one of the interpreters out there, have him say it into the voice recorder, rename the file, and you're good to go (using the voice recorder interface, not the HTML interface).

For instance, if you're on a mission and you find that you need to search for Hassan Ahmed and you have the Phraselator, you're screwed if you want to use the Phraselator interface. We've had to do numerous additions in the field on the fly for special circumstances, and we were able to do so quickly and effectively using the built-in voice recorder; no purchased software required!

3. One of the arguments for the Phraselator is that the translation is "hands free," since it is speach to speach. Thus, in theory, a soldier will be safer as they can keep their hand on their weapon while translating. This is a GREAT example of someone developing an operational need sitting on their ass behind a desk in the US that has no basis in military operational reality (although I'm sure their heart is in the right place). The operational reality can be summarized as follows: There are ALWAYS two soldiers pulling security when a third is communicating with a local. When you are communicating a complex set of phrases, you do not want your hand on your weapon for intimidation reasons anyway. And lastly, if the situation is so tense that all soldiers present need to have their hands on their weapons, trust me, there is nothing complex that needs to be said. This operational truth has been discussed extensively with soldiers in combat, and has been proven out on the all missions (numerous) that I've been on with them at the front lines.

[EXCEPTION: There is an exception to the last statement, generally in the case of HA (Humanitarian Aid) missions and medical facilities. These operations generally need to be able to say a lot of different things when communicating with patients. The speach to speach interface is great here, as it effectively sorts through a long list of items/phrases. In our case, we don't care about that sorting functionality as we mainly use only 1-3 phrases in house and village searches, and about 7 phrases at Traffic Check Points. The Charlie Med facility here on base has been using the iPaq, love it, and use it every day for translation, but I've recommended to them to evaluate the Phraselator as I think it will much better suit their needs.]

4. Doing simple voice recordings on the iPaq is effectively free. There's no software to buy. Plus, we're able to do interesting stuff easily, like drop everything into an HTML interface that is consistant with photos, lists, 9 Line Medivac info, etc. The Phraselator software is an added expense that is completely unnecessary, especially in light of the fact that, as noted above, the prerecorded phrases to date are useless out here, and there is no operational needs basis for speach to speach.

Now, the one thing that I do REALLY like about the Phraselator is the hardware option. It's a semi-ruggedized PPC with a louder speaker than the standard iPaq (not necessary, but nice), and most importantly, it takes AA batteries directly, while with an iPaq we have to either recharge off AA batteries or a BA-5590 (mil battery); added tasks that the soldiers could do without. The DoD has just gone through with a purchase of about 2,000 Phraselators, which is why they've been in the news so much. However, these ruggedized Phraselators cost somewhere between $2-3,000 EACH, whereas you can get an iPaq 3835 for something like $300 each if you want to go super cheap. (More for hard case and charging equip, but that can be leveraged later with different iPaqs when you want to upgrade.)

Phraselator's website claims that Phraselators are being used in Afghanistan for operations. Yeah, right. I'm here. Show me one. The Phraselator guy was even supposed to come pay us a visit two weeks ago. Where is he?

Believe me, we considered the Phraselator before deciding to go with a simple iPaq solution; the people who are funding us needed good reason to go against what the DoD is going to buy. However, since we don't like the Phraselator software, paying $2K+ for some nice hardware and then loading the phrases we actually care about into the voice recorder just wasn't worth it.

The thing that cracks me up about how we're just using the built in voice recorder as a translation device is that this could have been done 2+ years ago but no one ever did. Or it could be done with a minidisc/MP3 player + small speaders, or a variety of other easy and cost-effective methods. There's really nothing particularly creative about what we're doing; but it works and works well. Now if I could just get the darn search function inserted in the PIE interface...

Make sense?

[Paul Martin]

kzemach,

Though my application was much less urgent from a military perspective, I had created a several pages of song lyrics I was trying to look through. For a while, I went a more complicated route. I installed Apache webserver and Perl on the device and found a free cgi script that would run. Honestly, it wasn't too bad, though not overly speedy. It did take up quite a bit of room as I don't know Perl that well and copied over all of the libraries to run it, probably too many. It took about 5 meg or so by the time I was done. To get things to work properly, setup takes a while, though if you had a "master" setup, you could export the appropriate registry settings and then import them on subsequent devices, using a registry editor. If you're interested, here's the site (http://www.rainer-keuchel.de/index.html). It's not real user friendly, but if your patient and somewhat computer savy, you should be able to get it to work.

That said, a much simplier solution might be to use Tascal's freebie search program (http://www2r.biglobe.ne.jp/~tascal/download/pocketpc/search_e.htm). It offers the ability to search for a text string. So, just browse to the directory containing the files, and give it a whirl. Since doing a hard reset on my device, I'm now just using the Tascal search. Perhaps I'll go back to the full webserver at some point, but only when I have more time to find a better solution.

Paul

[kzemach]

Paul,

Thanks! I'll check out both those options as well as the javascript one. What I really want to know is how did anybody accomplish ANYTHING before the internet?

Thanks again for the leads,

Ken

[Paul Martin]

Ken,

I really wonder. It's the first place I turn for information, advice, you name it....especially when you're the only "techie" you know. :^)

By the way, Rainer (the Apache/Perl guy) has a discussion group on yahoo groups that is not a bad place to look through for advice (not to friendly to noobs, however IMHO). Look for [wince-devel].

Paul

[Pony99CA]

Thanks! I'll check out both those options as well as the javascript one. What I really want to know is how did anybody accomplish ANYTHING before the internet?
I doubt the Tascal Search will meet your criteria, if I understand them correctly. You don't want the soldiers to have to switch out of your translation HTML page to find a phrase, right?

I can see two ways around this. The first is to give each translated audio file a name that corresponds to what is being said in English, then just do a wildcard file search on those names. A good search program will allow you to double-tap the file to play the translation. If you just use the Voice Recorder to record the notes, the built-in Pocket PC Find program will do that. With this solution, you might not even need the browser.

The downside is that you'll likely have long file names, maybe even longer than whatever the Pocket PC name limit is.

The second thing you might try is getting the Pocket Internet Explorer plugins to support a Find function. I've often thought that would be useful, but I haven't really had the need to request it. Maybe you could send a request to the authors of MultiIE (http://www.southwaycorp.net) and Spb Pocket Plus (http://www.softspb.com/products/pocketplus/index.html).

The downside is that you'll be depending on someone else to implement this, and their priorities may not match yours. Offering them some money to make the addition might get it done more quickly, though. :-)

Also, Jason is the VP of Marketing for Spb, and maybe they'd be interested in being able to claim their software is being used to help U.S. troops in Afghanistan. <hint, hint> :-D

Steve

[humbletim]

hi ken--

here is a short PIE HTML/JavaScript example that 'hilites' in-document searches (entirely client-side).

you would have to pre-process your HTML-saved Excel spreadsheet through a "javascript'er" (i wrote a Perl script to do this). depending on how large your documents are, this may be a viable zero-install solution.

this approach could be modified to use anchor tags to simulate 'find next' searching.

i have tested this with IE on a T-Mobile Pocket PC Phone Edition (ARM processor).

if anyone wants more detail, feel free to email me.

later,
--tim

<html><body>
JavaScript hilite demo -- tim at etewireless dot com<br>
<form method="get">
<input name="search" type="text"><input type="submit">
</form>

<script language='javascript'>
var tmp = document.location.href;
var search = '';

if (tmp.indexOf('search=') >= 0) {
tmp = tmp.substring(tmp.indexOf('search=')+'search='.length);
tmp = tmp.substring(0,(tmp.indexOf('&') > 0 ? tmp.indexOf('&') : tmp.indexOf('#') > 0 ? tmp.indexOf('#') : tmp.length));
search = tmp.replace('+',' ');
}

var str = '';

str += '<p>\nIs there a plug in or upgrade (I\'ve looked and';
str += ' can\'t find any) for Pocket IE that would allow on';
str += 'e to search within a webpage for text, like you ca';
str += 'n in regular IE by doing a &ctr&F. \nWe coul';
str += 'd really use that functionality for searching for ';
str += 'information on our units. The webpages are all off';
str += '-line (on the actual iPaqs); perhaps there\'s some ';
str += 'way we can embed an internal search function in th';
str += 'e HTML??? \nI\'m not optimistic, but if anyone know';
str += 's of an upgrade or way to do this, it\'d be great. ';
str += '\nThanks! \nKen\n</p>\n';
str += '';

if (search) {
var strtmp = '';
while (""+strtmp != ""+str) {
strtmp = str;
str = str.replace(' '+search,' <font color=red><i><b>'+search.substring(0,1)+'</b><b>'+search.substring(1)+'</b></i></font>');
}
}
document.write(str);
</script>
</body></html>

[kzemach]

That's aweome! Im in Germany right now and havn't been able to try it yet, but will when I return. Should work, all the documents are relativelz small.

Thanks everyone,

Ken

[pivo]

Hi,
You might want to check Pocket Voice (http://www.pocketvoice.com). Although it was designed to help people with speech impairement (this is the main reason why the user can choose between Text, Symbols or Images as an interface) it has been used with foreigner languages. Last year during the World Soccer Championsiph held in Korea and Japan some friends of mine had about 100 phrases recorded (they found some people at oriental restaurants willing to help) before travelling. They told me they manage to 'live' during the 2 weeks they were there even most people didn't undestand any language but their own.
The English version is finish although we're checking grammar and spelling and will be made available in some weeks. At this time beta testers (for the translation only because the Portuguese version, that is free, is reaching the 2000 users).
Pedro