SMS Poll

[jducoeur]

Okay, here's an odd-sounding question for all those of you who use SMS, aka text messaging from your phone -- specifically, those who have ordinary cell phones with normal number pads, not full keyboards. (Those of us with Treos are assumed to not be the usual case.)

We're trying to come up with a concise and easy-to-type SMS syntax -- which means I need to check out what's actually easy to type. So I'd like to know what's under your buttons. If you could take a minute or two to look at your phone and answer this, I'd appreciate it.

I think of the "common" mobile-phone letter layout as being:

• 2: abc


• 3: def


• 4: ghi


• 5: jkl


• 6: mno


• 7: pqrs


• 8: tuv


• 9: wxyz

[Poll #1089759]

Comments

[lady-guenievre.livejournal.com]

My answers may be slightly off "standard" because I have Sprint's "intuitive" texting turned on - they call it "Word English" it's supposed to figure out what I'm saying and fix it without me having to hit letters twice - ie I can hit 8398 and it'll figure out I mean "text" instead of "tdwt", etc. Punctuation is much the same... It's a Razr, and I suppose they're pretty common - certainly something like 50% of my friends have one.

Predictive text

[metageek.livejournal.com]

Yeah, this is actually pretty standard these days. I'd be surprised to see any system that doesn't support predictive text getting any significant traction.

Re: Predictive text

[lady-guenievre.livejournal.com]

True, just wasn't sure whether that would throw off the poll...

Re: Predictive text

[metageek.livejournal.com]

Yeah, the split between raw typing (of any sort) and predictive text makes it hard to come up with a syntax that's easy for everybody.

[siderea]

Oh. I have a Razr. I didn't know that. Huh.

[jducoeur]

Good point. I have no idea how predictive texting is likely to screw us up, but it's something worth thinking about...

That's the standard touch-tone layout

[metageek.livejournal.com]

You won't find anybody who doesn't have that layout of letters. It's derived from the standard touch-tone layout (dating back to 1960); the only change is adding Q and X in the obvious places.

Re: That's the standard touch-tone layout

[jducoeur]

That's fine. I knew that it was common, but it's important to know that it's approximately universal. (At issue is a notion of allowing people to substitute the appropriate letters for numeric codes in one place -- this makes sense iff the correspondence is consistent.)

[metahacker.livejournal.com]

This is a phone from like 2001, though.

[yukirien.livejournal.com]

My answers were given by a Motorola slide phone. The phone I used for texting in France was a bit different in what 1, #, *, and 0 would do. I remembered never using 1 or 0, because it didn't really do anything important. # controlled all changes in capitalization and to punctuation and to numbers while * gave spaces. I never use texting in the US though, but I also never really use my cell phone at all in the US.

[baavgai.livejournal.com]

Just a side thought. I always thought if I could have an alternate entry method to the semi standard keyboard described, I'd me much happier.

What if only 1 2 and 3 worked? I could easily hover my thumb over three buttons without looking. My starter options are something like

1:A-I, 2:J-R, 3:S-" "

I choose 2 and get another menu
1:J-L, 2:M-O, 3:P-R

I choose 2 again and get
1:M, 2:N, 3:O

And 1 get me an M.

Every letter and a space are just three clicks of three keys away. (3^3).

Sure, it looks kind of complex, but it's no more of a learning curve for me than trying to always figure out the little letters under the little numbers.

[jducoeur]

Hmm. It's largely a wash to me, but I'll point out that it does involve slightly more thumb movement, since a given character may involve three different buttons; by contrast, the "usual" model is all a single button, although possibly pushed multiple times. On the up side, it involves less travel for the thumb, so it is probably easier to do without really looking. Possibly on the down side, it involves fewer letters than can be achieved by a single keypress, and the already-dicey predictive text systems completely can't work with it...

[baavgai.livejournal.com]

Yes, it's all thumb movement, just locked into the span of three buttons and the same horizontal line. The movement itself is relatively low stress, particularly compared to all the other movement involved in changing vertical orientation. In the layout I described, any letter will always involve three presses. It would never be few presses, just less movement. Out of curiosity, I worried out what a message would take.

The message "phone type
" in the traditional "7-44-666-66-33-00-8-999-7-33
" and the three press "231-132-223-222-122-333-312-331-231-122
". Yep, it's longer, but look at all the up down on the traditional. Makes my carpometacarpal joint itchy just looking at all those vertical transitions in the standard method.

Just thought to would amuse and might help with what you were thinking. Only half serious, though if such an option were on my phone I'd definitely give it a go.

[jducoeur]

Oh, I understand. I'm mostly treating it as an intellectual exercise -- it's an interesting idea, with some pros and cons balancing each other.

It's not actually relevant to what I'm doing -- we're focused on the *numbers*, not the letters. Under some circumstances, the user will need to type a short numeric code; the question is whether it is practical for us to allow them to type letters instead, to make it quicker. So if someone types "ad", I know that that's actually code "23" -- the main point of the poll is to check whether I can reliably count on that mapping. But the numbers are the point, so we're going to be using most of the keypad. (The codes are short, though, so I don't expect that to be a big issue.)

[johno.livejournal.com]

Moto Razr