By Rhodri MarsdenI got a new mobile phone on Wednesday. (It was a Nokia N95, in case a certain Finnish mobile phone company want to reward me for making public my brand loyalty.)
Of course, the day was immediately written off in an extended attempt to make the new phone behave exactly the same as my old one did. You become comfortable with a phone, you spend a year or more honing all its little intricacies to fit seamlessly with the cut and thrust of your life. And then you're given a new lump of plastic that you need to spend precious time stamping your identity on.
The most irritating manifestation of this crops up in the world of predictive texting, the feature which auto-completes words as you're tapping out SMS messages, supposedly saving you time and effort. But Nokia's dictionary doesn't contain certain words that are absolutely essential to everyday communication – like, for example, Rhodri. And Tooting. And Cyberclinic. So you have to teach your phone to recognise these words in order to save you having to laboriously spell them out each time.
But no brand of mobile phone – as far as I'm aware – allows you to export your custom dictionary from your old phone and import it to your new one; you have to start teaching it all over again, from scratch. It's like bringing up a child, I imagine. Except with fewer nappies.
Anyway, after having to teach perhaps my eighth mobile phone how to spell "Rhodri" (it invariably wants to call me "Shoesi" or "Pinesi") I thought that's it, enough is enough. And with a bit of sleuthing, I managed to work out how to salvage my old phone's custom word list and get my new phone to start using it. This trick might only work for people with modern Nokia phones and leave the rest of you gnawing off your fingers in frustration; I'm sorry about that.
Using a utility called Y-Browser – which is just a more powerful File Manager and lets you see parts of the phone that Nokia wouldn't normally let you see – I found on the C: drive of both phones a folder called "/Predic/". And, inside, a file called "101F8615_F.dat". I copied the old "101F8615_F.dat", over to the new phone, overwriting the new "101F8615_F.dat", and bingo. Suddenly my phone recognises my christian name. We're hitting it off. We're best pals.
What's marginally more amusing is the actual contents of "101F8615_F.dat". I opened it up in a text editor on my computer; and stared in horror at the words that peeked out in between all the code. People who know me and who exchange text messages with me may find the words contained within this file somewhat familiar. People who don't know me might imagine that I'm a foul-mouthed drunkard. But I'm not. Honestly.
CONFUSED ABOUT TECHNOLOGY? SUBMIT YOUR QUERIES TO CYBERCLINIC USING THE COMMENT FORM BELOW, OR EMAIL QUESTIONS HERE.
Presumably that means you could easily remove all the misspelled words that make their way into predictive text too?
Posted by: John | Friday, 02 May 2008 at 02:06 PM
I don't think so, John. While it's easy to do the old switcheroo with the files, actually getting in there and editing the content of them is a different matter. There's invisible characters & code running throughout the file. Someone with more of a programming background might be able to attack it, but not me.
Posted by: Rhodri Marsden | Friday, 02 May 2008 at 02:17 PM
Cool post.
I got excited for a sec, when I saw your text file.
I thought I had found what i was looking for!
I have to get my head around the rules used for predictive text.
As a starting point, I am trying to get my hands on a simple text file that lists all the words that are part of the default mobile phone dictionary.
It seems quite hard to find.
Is it known to be the same as, for example, the oxford dictionary.
Is there a dictionary file that rates the words into a "predictive text word order", so that common words appear before obscure words.
ie. What makes "Alot" appear before "Clov".
I have an interesting task I am trying to do, to make sure that a text code I give people is the first one that predictive text will display ....
I am trying to generate a list of words that "appear first" (based on the default predictive text rules in the phone), based on every number combination.
For example if I type 2568 into my mobile phone, the first word it predicts is "Alot"
followed by this list in this order:
Blot
Clou
Alou
Blou
Clov
For my project I am only interested in the first word that appears (eg. in this case "Alot") when you type any set of numbers.
I am a but stumped as to how i will generate this list for all number combination's.
eg:
2333
2334
2335
2336
2337
etc,etc,etc,etc,etc,etc,etc,etc,etc,
I guess i need a list of all valid words, grouped based on the number combination of the first 4 letters, then rated the same way the predictive text does.
Any ideas where I could get this text list?
I hope you dont mind me posting this question here.
I am just a bit stumped as to how to approach this task.
Thanks heaps for your blog,
Mark
Posted by: markv | Tuesday, 02 September 2008 at 02:12 PM