[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

SEUL: Part Two



On Mon, 5 Jan 1998, Paul Anderson wrote:

> What about LyX?  Wordperfect?  Hasn't CorelDRAW been ported, too? 
> xcalc, xload, date.....  Lynx?  Arena?(Blech) xtrek? Quake?  DOOM? 

Wordperfect has a Java version, as does CorelStuff (which is wordperfect
now).  They have a native linux too.  On the other hand... they're not
exactly top caliber or up to date.  I get the feeling that they are
something that the employees did in their spare time.  Lynx is heading off
into irrelevance, as I touched on before, because it increasingly can't
support the modern world.  No frames, no java/javascript, no tables, and
it can't display a page while loading.  If Netscape were programmable,
Lynx would be finished except for browsing HOWTO's. 

Arena blows, xcalc and date hardly count, and xtrek isn't exactly Ultima
Online.

The only company that supports Linux with its heart and soul is id
software, makers of Quake and Doom.  On the other hand, that's one modern
game and one aging game, plus a smattering of trinkets like Angband,
nethack, xgalaga, and xtrek.  There's thousands of these for windows and
hundreds of professional quality games.  Even the third parties that
license the doom and quake source don't bother to port it to linux, even
though it would be trivial.

> > 2) We need better fault tolerance.  Installations in particular.  Red
> > Hat is as guilty here as anybody if not more.
...
> Not any more, they're improving.  Ibid.
 
Well, whatever.  I still haven't had a Red Hat install go flawlessly the
first time, except with the Kickstart disk.  It always crashes.  I can say
that with confidence.  Always.

> > There will be users who will just turn off the computer.
> > They can't hose their data in the process.
...
> Same prob with Win95, it's called "disk caching".  I trust you remember
> what it is...  Speeds things up more than you might realize. 

Not hardly.  Disk caching is one thing, and having your filesystem corrupt
if you lose power is quite another.  Granted linux is less susceptible to
this than most Unices, and Win95 has the same problem.  On the other hand,
with Win95 all you ever have is some lost clusters.  Digging everything in
/home out of lost+found is quite another. 

> Maybe RedHat missed something, 'cause they've been distributing the JDK
> with their distrib since 4.2. 

No, that's kaffe and guavac.  Which are similar, but, as I mentioned, they
don't work in 5.0 (and the fix is nontrivial).  As long as you can count
on companies NOT distributing a fully tested JRE with their Java programs,
simple things like this HAVE to work. 

> > 4) We need better dial-up networking.  If Linux wants to have any
> > chance in the home market, or even in most of the business market,
> > there has to be a click-and-play dial-up networking. 
>
> It depends on whether the ISP supports PAP or not.
> If they do, just do as such:
> Creat /root/chat, with
> "" ATZ OK ATDT8888888 CONNECT
> (Replacing 8888888 with your ISPs number)
> Then have an /etc/ppp/options like thus:
> /dev/cua1
> defaultroute
> connect "chat -f /root/chat"
> lock
> 115200
> +ua /etc/ppp/pap-secrets
> 
> And put your username and password in /etc/ppp/pap-secrets.  Simple, no?

Frankly, no.  With win95 you click on "create dial-up connection," select
a modem, and tell it the phone number.  This is certainly not a
two-clicks-and-a-number process.  It requires the user to know how to
create configuration files, to understand device files, to understand the
syntax for a chat script, and to know about things like bit rate and
TCP/IP.

> > Ideally, a system that could take a windows' DUN-script and translate
> > it would be ideal.  Not too hard, either.  Could probably be done in a
> > day with perl. 
> 
> Not really...  IIRC, Win uses PAP, right?  Grabbing the password and
> username from a doubtless encrypted file with PERL?

No, the user can be expected to remember their password.  I don't think
Win95 encrypts it anyway.  What I am referring to is the scripts that
people like Compuserve distribute with their internet packages.  It takes
the user through Compuserve's maze of services to get to PPP, then
initiates the PPP connection.

> >  And if you pick 'install everything' it breaks. 
>
> Not in my experience, AAMOF a friend of mine(Linux newbie, but not a
> newbie to computing) installed RH4.2, selected "Install Everything" and
> he didn't have a single problem.

In Redhat 4.2, "install everything" would cause any Novell network and a 
substantial percentage of Windows Networks on your ethernet to break.
Granted, this could be considered a problem with the design of the
protocols or servers... but still.  It happens, and it can't.

In Redhat 5.0, "Install everything" will foul up your routing so that it
is unusable.  I think it still breaks Novell networks.  (I don't know, I
don't have Novell on my network anymore).

> Ha ha, what kind of computer you talking about?  Win95 install takes AT
> LEAST 15 minutes, with lots of baby sitting...  RedHat has recently come
> up with an installer called "KickStart", shove the boot disk in A:, put
> the CD in the drive, reboot and walk away, not even Win95 can boast

Well, I was talking about selecting the packages, not the complete
install.  Kickstart is a great thing that Redhat has invented, and I
heartily support it.  It's the exact sort of thing we need.

> Okay, now for my one gripe:  XFree86.

A little over two years ago I spent twelve hours configuring XFree and am
still using the same configuration (and the same monitor and video card). 

This is the sort of thing that would NEVER work in Windows.  :)

> sendmail, and we're talking direct edits of sendmail.cf).  THE problem 
> is vertical and horizontal frequencies.  The XF86Config file comments
> recommend grabbing the appropriate numbers from your monitor's manual.
> First, I have yet to see a monitor manual that mentions this, and

Sony, Mag and NEC manuals at least that I have seen document this.  NEC
gets the numbers wrong, though.  :) 

Without help from the manufacturers, the best way to do this is to
EXTENSIVELY recruit updated information from modern linux users.  Finally,
like you suggested, do as Windows does - test it and see if it works.

:)