It would be nice if there a light version of DCS.

DCS could be optimized down to one page (an no Easter eggs )...

Other ideas:
- An option to block the memory menu (or require a password).
- Be able to start other programs up at start-up ( including archived stuff ASM programs that aren't no-stub ).

The Easter Egg has already been shrunk from 3k to about 300 bytes by removing the entire audio sequence. Guys, debate this. From a technical standpoint it's nearly impossible to find all the libraries into 16k (and may well be full-blown totally impossible) but I'd be interested to hearing what you guys think. I'm of the impression that for the number of people who would actually end up using this 1-page version, it's simply not worth it to go through all the trouble of trying to fit it in that space.

No, I wouldn't use a DCS Lite (though it would be nice for those, eh, poor 83+ users). Quite the contrary, let's turn this into 3 app-pages!
Anyhow, you could, for starters, take out support for SE's and take out the CalcNet thing, I don't know how much you'd save there. Or, you COULD just leave it to brandonw to finish up Noshell, but you wouldn't want THAT, now would ya?
But seriously, integrated pw protection is a MUST. And (jumping to 6.2) how about execution of Ion, MOS, and DCS ASM prgms from within BASIC prgms (or is that what No----l is for? )

The only way it would be remotely suitable for a 3 page shell, is if it offered enough libraries to make for ungodly small program sizes. And mabye a 'DCS lite' could be a version of dcs that didnt have MOS routines (or mabye ION if they take up a lot of space) so as you could play DCS games only on DCS lite.

But DCS developers develop knowing they can use MOS and ION routines, so that wouldn't work..

Kerm, how large would dcs 'lite' be if you took out MOS routines and ION routines, if small enough, you could add extra replacement DCS routines.

There aren't many ION routines, and they are heavily used - why would you want to remove them???

Besides, there are more pointless things that can be removed for a "lite" version - such as the calcnet routines.

I suggested removing calcnet to Kerm, but thats why I asked the size of the Ion routines, because I know they are heavily used.

MOS routines take 4-5kb, Ion routines take about 1k. There's _no_ way I'm removing the Ion routines. Also, CALCnet2 only takes about 600 bytes.

How big are the DCS routines

Don't forget the GUI.

Which is why Noshell could end up being quite economical for those w/ very limited memory.

The GUI API routines are perhaps 12k all together. The program running support is about 5-6k total, and the rest of the interface must be 3-5k.

I think a DCS 'lite' version would generate a lot of anger from n00bs. Like, they download it, but then can't run their stuff 'cause they're n00bs, and then they flame Kerm about it everywhere. I think just one version is good, it reduces the amount of documentation needed, and will keep people from getting angry. If they want a one-app-page shell, they should use MirageOS, and just live with the lack of cool GUI. That's just my opinion, though, so I'm not saying Kerm shouldn't make a DCS 'lite', I just don't think it would be very wise.

If it's possible to keep it in 1 page without removing any routines (with the exception of CN2/GCN, those could be removed), I say go for it and just strip out most of the shell's actual GUI; just make it run programs.

So make it like noshell?

I agree, though I don't think Cn2/GCn needs to be removed in DCS Lite because Kerm said he will try to make some demos of Cn2 in action over the summer. As for GCn, I don't know when it's going to be online, but imagine a calculator MMO game, with every connected via GCn...

Exactly, but without hooks. Do something like the ION GUI; keep it small and simple

I'm not sure what is going to happen with GCN. At this point that is more up to Kerm to figure out how he wants to make the calc<->comp interface. Initially he didn't want a separate protocol and wanted the computer to act like a calculator on the network, but that requires either using an old link cable (such as a parallel port) and custom drivers, or the creation of a custom interface/controller (such as the USB PIC I made but never coded to work with Cn2)

As I've said several times, the CALCnet routines take up a grand total of around 400 bytes. At this point, a DCS Lite edition is not technically feasible, although it may be at some point in the future. For now, to save confusion, I'm going to stick with a single branch.