RE: HW requirements

RE: HW requirements

[Feneric Brown]

> I really appreciate you doing the packed Boolean tables; it makes a big
> difference to people with less memory.  (maybe we should think about
> Xconq on PalmPilot or Cellphones?   Games on those things seem pretty
> popular  ;>  )

Actually porting Cconq to PalmPilots, Newtons, Psions, Nokia GEOS 
units, etc. may not be that awful.  The biggest issue is probably 
getting the build kit / SDKs / etc. for each; I don't think all of them 
offer free development environments (although some definitely do).  Of 
course, Cconq itself can't network with a proper Xconq game at the 
moment, so that might limit the usefulness of such ports.

RE: HW requirements

[Eric W. Brown]

> Can the serial port on Commodore 128 even keep up with a 56K
> modem?

The misleading (but honest answer) is "not for long".  It (the whole 
system including serial port and software emulated UART) can handle up 
to 9600 bps without too much trouble (the C64 only up to a little over 
2400 bps) but not the full bandwidth of what a modern modem can 
deliver.

The more accurate answer however is "it doesn't have to".  Virtually 
everyone who's still using a C128 (or C64 for that matter) these days 
connect not through the native serial port, but rather through a UART 
cartridge.  There were at least three such cartridges on the market for 
fairly low money (they've always tended to be cheaper than the modems 
themselves), but competent hardware hackers have also been known to 
build them using freely available plans (do a search for "DataPump" or 
"DataPump Plus" in conjunction with "C128" or "C64", and I'm sure 
you'll be able to find a copy somewhere -- there used to be a bevy of 
FTP sites that made it available for anonymous download).

> And since I feel a wave of nostalgia coming on, __I wonder how
> easy/hard it would be to make Cconq into a BBS "door"....

That's an interesting question.  If the BBS were also to support a web 
front end, it could make for a unique Xconq experience...

Re: HW requirements

[Bruno Boettcher]

On Sun, Jan 11, 2004 at 09:11:10AM -0800, Erik Jessen wrote:
hello!

ok i followed this trail with mild interest, but have this time to step
in myself :D

> Thanks!  I know people who do ASCII-only for email, etc. because they
as me...

> either have a slow connection, or want to avoid viruses.  I'd expected
uhm also commodity? as Feneric pointed out i am one of those users that
mangle incoming mail through numerous utilities, and i never found any
other MUA that is as practical and fast to use and not to say easy to
propagate through the net (means i want to consult my mails in usual way
from everywhere i happen to be) as mutt...

> that for playing Xconq, they'd have to have newer hardware (and/or a
> fast connection) to play Xconq, simply because of RAM/CPU
> considerations.

>> 4.  Sophisticated users who choose to route all their e-mail through a 
>> system including combinations of things like Procmail, SpamAssassin, 
>> JunkFilter, TMDA, Pine, Mutt, etc. in order to largely sidestep the 
>> issues of spam and virii that so plague the modern world.
>> 
>> We maintain a bunch of games on the system, and I'd have to say that 
>> Cconq (the VT100 version of Xconq) remains one of the most popular, 
>> right up there with NetHack and the various IF titles.

uhm i admit i am too a regulare ascii nethack player :D but in the case
of xconq i use the tcl/tk interface :D
and i have allways up to date machines, since i do number crunching for
my living.... which doesn't mean i have unlimited ressources, since the
main ressources go for the simulators running, which allready caused
several times the killing of some xconq games for outrunning
ressources...

-- 
ciao bboett
==============================================================
bboett@adlp.org
http://inforezo.u-strasbg.fr/~bboett
===============================================================

RE: HW requirements

[Eric McDonald]

Hi Eric, others,

On Sun, 11 Jan 2004, Eric W.Brown wrote:

> Oops, meant to say "it's possible to play Cconq on a remote server with 
> a C128 connected only with a 56K modem."

Can the serial port on Commodore 128 even keep up with a 56K 
modem?

And since I feel a wave of nostalgia coming on, __I wonder how 
easy/hard it would be to make Cconq into a BBS "door"....

Eric

Re: HW requirements

[Jim Kingdon]

> That said, it is hard for me to "feel the pain" when it comes to 
> memory crunches, as the machine I use for development has 1 giga_, 

Just look at how much memory is shown for the xconq process in "ps".
Granted, you have to bother to check....

> I could lie to the Linux kernel about how much memory I have by giving
> it a mem=64 argument from the boot loader or something....

I suppose this would the best way of getting an idea of how badly
paging affects xconq's performance.  I suppose one could try to have
some other process allocate a bunch of memory and keep accessing it
(which I suppose in some ways better resembles many real-world
situations), but the mem=64 thing is simpler.

Re: HW requirements

[Jim Kingdon]

> Actually, as long as the host machine is beefy enough, the people 
> connected can play Cconq with much lesser clients.

Just in case it was unclear to anyone on first reading, "client" and
"host" here refer to remote login (say, via ssh or dial-up or
whatever).  The xconq protocol would presumably require roughly the
same resources on the host machine or the joining machine, since they
both keep a complete copy of the state and do all the calculations in
parallel.

RE: HW requirements

[Eric W. Brown]

> I'm pretty sure it's possible to play Cconq on a remote server while 
> connected only with a 56K modem.

Oops, meant to say "it's possible to play Cconq on a remote server with 
a C128 connected only with a 56K modem."

RE: HW requirements

[Feneric Brown]

> Thanks!  I know people who do ASCII-only for email, etc. because they
> either have a slow connection, or want to avoid viruses.  I'd expected
> that for playing Xconq, they'd have to have newer hardware (and/or a
> fast connection) to play Xconq, simply because of RAM/CPU
> considerations.

Actually, as long as the host machine is beefy enough, the people 
connected can play Cconq with much lesser clients.  Although I've not 
tried it myself (now that I've said it, you know I have to!) I'm pretty 
sure it's possible to play Cconq on a remote server while connected 
only with a 56K modem.  Connecting from an old SLC via ethernet should 
be blinding by comparison...

RE: HW requirements

[Erik Jessen]

Feneric,
Thanks!  I know people who do ASCII-only for email, etc. because they
either have a slow connection, or want to avoid viruses.  I'd expected
that for playing Xconq, they'd have to have newer hardware (and/or a
fast connection) to play Xconq, simply because of RAM/CPU
considerations.

This is really interesting.

Erik

-----Original Message-----
From: xconq7-owner@sources.redhat.com
[mailto:xconq7-owner@sources.redhat.com] On Behalf Of Feneric Brown
Sent: Sunday, January 11, 2004 8:48 AM
To: xconq7@sources.redhat.com
Subject: RE: HW requirements

> How many people use VT100 any more?

You'd be surprised.  I do the maintenance at a company that provides a 
number of shell accounts to customers.  The users typically fall into 
one or more of the following groups:

1.  Sophisticated users / designers who want a decent UNIX-like 
environment with fairly up-to-date tools without having to maintain it 
or handle the updating themselves;

2.  Users who are blind and have all output redirected through either a 
reader box or a Braille hand reader;

3.  Users who are opposed to the upgrade cycle and who (often for 
philosophical as well as financial reasons) have chosen to freeze their 
hardware at a certain level that handles their own requirements (the 
C128 is popular, but there are other makes & models  that are also 
used);

4.  Sophisticated users who choose to route all their e-mail through a 
system including combinations of things like Procmail, SpamAssassin, 
JunkFilter, TMDA, Pine, Mutt, etc. in order to largely sidestep the 
issues of spam and virii that so plague the modern world.

We maintain a bunch of games on the system, and I'd have to say that 
Cconq (the VT100 version of Xconq) remains one of the most popular, 
right up there with NetHack and the various IF titles.

RE: HW requirements

[Feneric Brown]

[-- Attachment #1: Type: text/plain, Size: 1254 bytes --]

> How many people use VT100 any more?

You'd be surprised.  I do the maintenance at a company that provides a 
number of shell accounts to customers.  The users typically fall into 
one or more of the following groups:

1.  Sophisticated users / designers who want a decent UNIX-like 
environment with fairly up-to-date tools without having to maintain it 
or handle the updating themselves;

2.  Users who are blind and have all output redirected through either a 
reader box or a Braille hand reader;

3.  Users who are opposed to the upgrade cycle and who (often for 
philosophical as well as financial reasons) have chosen to freeze their 
hardware at a certain level that handles their own requirements (the 
C128 is popular, but there are other makes & models  that are also 
used);

4.  Sophisticated users who choose to route all their e-mail through a 
system including combinations of things like Procmail, SpamAssassin, 
JunkFilter, TMDA, Pine, Mutt, etc. in order to largely sidestep the 
issues of spam and virii that so plague the modern world.

We maintain a bunch of games on the system, and I'd have to say that 
Cconq (the VT100 version of Xconq) remains one of the most popular, 
right up there with NetHack and the various IF titles.

[-- Attachment #2: smime.p7s --]
[-- Type: application/pkcs7-signature, Size: 2365 bytes --]

RE: HW requirements

[Erik Jessen]

My home machines all have at least 512MB RAM.
(have to - my kid's games want at least that much...  I should go write
kid's games that suck up a lot of RAM, and buy stock in Micron, so I
make
all my money off the memory that people will buy to play the games ;)

At work we use 1-12GB RAM, but that's because we have to.

I really appreciate you doing the packed Boolean tables; it makes a big
difference to people with less memory.  (maybe we should think about
Xconq on PalmPilot or Cellphones?   Games on those things seem pretty
popular  ;>  )

Regards,
Erik

-----Original Message-----
From: Eric McDonald [mailto:mcdonald@phy.cmich.edu] 
Sent: Saturday, January 10, 2004 11:45 PM
To: Erik Jessen
Cc: 'Jim Kingdon'; xconq7@sources.redhat.com
Subject: RE: HW requirements

On Sat, 10 Jan 2004, Erik Jessen wrote:

> So, is anybody even really considering HW resources when actually
doing
> development?  (other than the general rule of "don't be wasteful")?

Well, that's the rule I follow. I try to be as conscious of memory 
and execution efficiency as possible without writing code that 
is too convoluted.

Since I may be adding some more tables of precomputed values soon, 
I recently implemented packed boolean tables. On machines with 
32-bit ints, this means using 1/32 the amount of memory that one 
might otherwise use with 1 bool per int. (Bit vectors can easily 
be created from this implementation, since they are simply 1 by n 
or n by 1 packed boolean tables.)

That said, it is hard for me to "feel the pain" when it comes to 
memory crunches, as the machine I use for development has 1 giga_, 
I mean, gibibyte of memory. I suppose, for testing purposes, I 
could lie to the Linux kernel about how much memory I have by 
giving it a mem=64 argument from the boot loader or something....

Eric

RE: HW requirements

[Eric McDonald]

On Sat, 10 Jan 2004, Erik Jessen wrote:

> So, is anybody even really considering HW resources when actually doing
> development?  (other than the general rule of "don't be wasteful")?

Well, that's the rule I follow. I try to be as conscious of memory 
and execution efficiency as possible without writing code that 
is too convoluted.

Since I may be adding some more tables of precomputed values soon, 
I recently implemented packed boolean tables. On machines with 
32-bit ints, this means using 1/32 the amount of memory that one 
might otherwise use with 1 bool per int. (Bit vectors can easily 
be created from this implementation, since they are simply 1 by n 
or n by 1 packed boolean tables.)

That said, it is hard for me to "feel the pain" when it comes to 
memory crunches, as the machine I use for development has 1 giga_, 
I mean, gibibyte of memory. I suppose, for testing purposes, I 
could lie to the Linux kernel about how much memory I have by 
giving it a mem=64 argument from the boot loader or something....

Eric

RE: HW requirements

[Erik Jessen]

Based on a few of the last emails, I heard:

a) 24MB is the min.
b) it's an unrealistic min.

So, is anybody even really considering HW resources when actually doing
development?  (other than the general rule of "don't be wasteful")?

I'm just trying to understand the environment Xconq is living in - I
simply don't have any data - I do chip design for a living, not
software.

Erik

-----Original Message-----
From: xconq7-owner@sources.redhat.com
[mailto:xconq7-owner@sources.redhat.com] On Behalf Of Jim Kingdon
Sent: Saturday, January 10, 2004 7:44 PM
To: xconq7@sources.redhat.com
Subject: Re: HW requirements

> How many people use VT100 any more?

One posted to the list just recently.

To me the real questions are not so much "how many users (or potential
users) with this or that hardware/OS/etc" but "is anyone willing to
test xconq on such-and-such setup?" and "is there any benefit to
assuming more memory/cpu/whatever?".  All too often (although
certainly not always) the answer to that last question turns out to be
"no", but one might only notice it when you've done some profiling
and/or other analysis about where the resources are going.

Re: HW requirements

[Jim Kingdon]

> How many people use VT100 any more?

One posted to the list just recently.

To me the real questions are not so much "how many users (or potential
users) with this or that hardware/OS/etc" but "is anyone willing to
test xconq on such-and-such setup?" and "is there any benefit to
assuming more memory/cpu/whatever?".  All too often (although
certainly not always) the answer to that last question turns out to be
"no", but one might only notice it when you've done some profiling
and/or other analysis about where the resources are going.

Re: HW requirements

[Hans Ronne]

>Hans Ronne writes:
> > As I said, 24 MB of RAM is required. I don't think disk space or CPU speed
> > is an issue on any machine that can run Xconq. The only serious limitation
> > on machines that people still use today is the poor graphics support in
> > pre-NT versions of Windows (98 etc). Which is a OS problem, not a hardware
> > problem.
>
>but there's still the curses version which does not require anything beyond
>a vt100 display

And between 3.5 and 8.5 MB of RAM depending on the game.

BTW, 24 MB is what you need to launch the standard game. It won't run long
on 24 MB, however. Some larger games (Ancient Near East) need up to 36 MB
to launch. Double that to make room for all the stuff that gets allocated
during a long game, and you get what Xconq really needs. For this reason, I
think the possibility of running Xconq on some of these older systems is
rather theoretical.

Hans

RE: HW requirements

[Erik Jessen]

How many people use VT100 any more?

Erik

-----Original Message-----
From: klaus schilling [mailto:510046470588-0001@t-online.de]
Sent: Saturday, January 10, 2004 4:00 PM
To: Hans Ronne
Cc: Erik Jessen; xconq7@sources.redhat.com
Subject: Re: HW requirements

Hans Ronne writes:
 > As I said, 24 MB of RAM is required. I don't think disk space or CPU
speed
 > is an issue on any machine that can run Xconq. The only serious
limitation
 > on machines that people still use today is the poor graphics support
in
 > pre-NT versions of Windows (98 etc). Which is a OS problem, not a
hardware
 > problem.

but there's still the curses version which does not require anything
beyond 
a vt100 display

Klaus Schilling

Re: HW requirements

[klaus schilling]

Hans Ronne writes:
 > As I said, 24 MB of RAM is required. I don't think disk space or CPU speed
 > is an issue on any machine that can run Xconq. The only serious limitation
 > on machines that people still use today is the poor graphics support in
 > pre-NT versions of Windows (98 etc). Which is a OS problem, not a hardware
 > problem.

but there's still the curses version which does not require anything beyond 
a vt100 display

Klaus Schilling

Re: HW requirements

[Hans Ronne]

>What about HW requirements (CPU speed, RAM, disk)?
>Where I'm going with this, is that machines typically have a 3-year
>life-span.  (before drives die, etc.).  So, figure 80-90% of customer
>base has a machine built since 2001.

As I said, 24 MB of RAM is required. I don't think disk space or CPU speed
is an issue on any machine that can run Xconq. The only serious limitation
on machines that people still use today is the poor graphics support in
pre-NT versions of Windows (98 etc). Which is a OS problem, not a hardware
problem.

As for the assumption that computers have a 3-year life-span, that may be
true in a corporate environment, but not elsewhere. I believe this has
already been discussed on the list.  We try to support a fairly broad range
of hardware and OS versions, not just the latest stuff.

Hans

HW requirements

[Erik Jessen]

What about HW requirements (CPU speed, RAM, disk)?
Where I'm going with this, is that machines typically have a 3-year
life-span.  (before drives die, etc.).  So, figure 80-90% of customer
base has a machine built since 2001.

Now, there will be people running older machines, but won't they also
want newer/faster CPUs, just to run the AI, and to have enough GDI to
run the program at all?

Now, I don't know what a mainstream machine was 3 years ago, but if
somebody could say, that'd help clarify just what kind of resources are
going to be available to players.

And thus, what kinds of resources the developers can count on being
present.

Erik

-----Original Message-----
From: xconq7-owner@sources.redhat.com
[mailto:xconq7-owner@sources.redhat.com] On Behalf Of Hans Ronne
Sent: Saturday, January 10, 2004 9:56 AM
To: Eric McDonald
Cc: xconq7@sources.redhat.com
Subject: RE: New Action: change-type

>On Fri, 9 Jan 2004, Erik Jessen wrote:
>
>> Just to ask, what's the min hardware requirements & OS we're
targeting?
>> One can buy a brand-new machine from Dell for about $300...
>
>Well, the minimum software specs we seem to be targetting is C89
>compliance. If a machine can build at least one of the Xconq
>interfaces with a C89 compiler and is running MacOS (with >= PPC
>processor), 32-bit Windows, or a fairly modern Unix, then it seems
>to be one that we are interested in supporting.
>
>Eric

In terms of OS's:

Windows NT, 2000 and XP - Fully supported.
Windows 98, SE and ME - Supported, but with reduced quality graphics*.
Windows 95 - Not supported (might still work).

* These older Windows versions have only 2MB of GDI resource memory.

MacOS 8.6 and above - Fully supported.
MacOS 8.1 to 8.5 - Fully supported native interface. No help system in
the
tcltk interface*.
MacOS 7.0 to 8.0 - Not supported (might still work).

* The obstack code and tcltk do not work well together under these older
MacOS versions.

Another limit is the amount of memory needed by Xconq (at least 24 MB).
This means that you are unlikely to be able to run it on machines with
ancient OS's.

Hans