Re: Marketing, was Re: C


12 Dec 1995 20:07:42 GMT

In article <4agl8a$cn1@magus.cs.utah.edu>, galt@spades.cs.utah.edu (Greg Alt) writes:
|> In article <4afpqu$ph3@senator-bedfellow.MIT.EDU> sanj@mit.edu (Sanjay S Vakil) writes:
|> >Hey there. I was the voice a while back that was trumpetting IF on the
|> >newton. It rocks. I mean it is *very* cool. The cost of the old MP100s
|> >is about $150 now, which is significantly more than the $40 you mention.
|>
|> Could you give some more info about the newton engine? And also about
|> the newton, as I know little. Like, how do you get the data onto it,
|> how much data can it hold, how long can it run on a set of batteries,
|> how much info can you see on the screen, can you scroll back to see past
|> moves, etc? Also, if someone wanted to tweak the code, how tricky would
|> that be? Would you need to buy a development package for the newton?

Not a lot to really tell at this point. It currently runs rather slowly -
newtonscript (NS) is the only language available to work on the newton, and
it isn't well suited to becoming a bytecode interpretter. It's based
on ZIP, with huge mods as required by a full language change (for example,
NS has *no* pointer math! argh!). In addition, the actual RAM is about
50-60k, the rest is functionally ROM.

The data is loaded by a helper program on a mac/unix/pc that loads the
datafile into a 'wrapper' package that can be loaded onto the newt, so
you still require a regular desktop machine. The newt itself, depending
on configuration, has 150k - 1.5M onboard user memory. You can get
RAM cards to up it by 2/4/8M chunks. My current machine has 150k of
onboard and a 2M flashRAM card. It runs on AAA batteries for about a
week (I use rechargeables). Newer machines run on AA batteries for about
a month. If you want to see a screen shot I took recently (for info that
can fit on the screen), look at:

http://web.mit.edu/sanj/www/home.html

I've already built in a 'ten screen' scrollback; that might be user settable
in the actual release.

Code tweaking will be a nightmare. I can barely keep up with it now, since
most of the speed hacks we're using make the code more complicated to
keep running. I don't think I'll bother making the source public, because
it won't help anyone, since you *will* still need the Newton Toolkit ($300)
to recompile it.

|>
|> >However, having a touch screen and never having to type is *wonderful*.
|> >
|> >Basically, the UI I've worked up lets the user tap on any word in the
|> >'output area' and have it appear as the next word in their 'input area'.
|>
|> But don't you still have to write most verbs? How about something that
|> keeps track of all the words that have been used previously. That way,
|> you would only have to write out each verb once (and the obvious ones
|> could be given already, like 'north,' 'south,' etc.).

Verbs are located on a 'picker diamond' that pops up a list of user settable
verbs. You can put stuff that you use often on that list, and can access
any verb in two taps. The directions are on a compass rose (the web page
snapshot makes this all very clear), with a few touch buttons for other
directions like "up", "down", "enter" and "exit".

|>
|> Greg
|> --
|> Videogames, Unicycling, and Anarchism: http://www.cs.utah.edu/~galt/

I feel like I'm creating all of this interest for a product that I won't
have time to finish until late January (I'm studying for qualifying exams).

I hope y'all will bear with me until then...
sanj