Subject:
|
Re: Perl rules!
|
Newsgroups:
|
lugnet.off-topic.geek
|
Date:
|
Thu, 22 Jul 1999 14:24:02 GMT
|
Viewed:
|
3753 times
|
| |
|
|
On Wed, 21 Jul 1999 23:43:27 GMT, "Don Heyse"
<donnell_heyse@adc.spam.go.away.com> wrote:
> In lugnet.off-topic.geek, Steve Bliss writes:
> > How did the segmented architecture enable the creation of tiny programs?
>
> Relative pointers in the same segment take up half as much space as a 32
> bit pointer. Depending on what you're doing, the savings here can be
> considerable. .COM programs used nothing but relative pointers and fit
> in less than 1 64K segment.
OK, relative-addressing mode is good. But it doesn't require segmented
memory. All it requires is an instruction format with a defined result.
For example, the conditional-jump instructions on the 6502
microprocessor[1] only used relative addressing. You could only jump in
the range of -128 to +127 bytes. So all the conditional jump instructions
took two bytes: one for the opcode, and another for the offset. But there
was no segmented memory addressing.
> That's tiny in my book. Nowadays, tiny
> rarely fits on a 1.44 meg floppy. Also the trend toward RISC architecture
> leads to code bloat for a lot of reasons that I'd have to look up. It's
> been a while since I thought about it. I sorta gave in and joined the
> bloat crew.
I know what you mean--I program in Visual Bloat.
> > > > Yeah, but what does that have to do with LDraw not supporting modes with
> > > > more than 4-bit color?
> > >
> > > Isn't 4 bit color the lowest common denominator between VGA and EGA?
> >
> > So? This is SVGA, and most video cards seem to not offer 4-bit color modes
> > in resolutions above 800x600. So LDraw will happily work in modes that
> > most video cards can't deal with, and LDraw won't do modes that video cards
> > will.
>
> Perhaps the original drivers were written to support only standard VGA and
> EGA and then were extended to support SVGA in compatible modes only.
> 640x480 x 4bits per pixel fits in one of those 64K segments and thus supports
> tighter code (A feature at the time)
But it wouldn't have cost LDraw much (if anything) to support higher
color-depths. It would just need to be able to talk to the library driving
the display. It wouldn't have to hold the higher-color bitmap itself.
> Hey, does LEDIT support any resolution other than 640x480?
Nope. WYSIWYG.[2]
> It seems more
> responsive to keystokes than LDAO/LDLITE right now, so I can work faster.
That's very true. I do much more parts-authoring than modeling, so I
mostly use LDAO/LDLITE. I need to look more at mlcad, to see if I can make
parts in that program...
> (I do like to ALT tab to the VEC though :-)) If only netscape had an all
> keyboard interface...
Doesn't it suck when programs don't have complete support from the
keyboard?[3] Especially high-cost development tools[4], with the layers of
dialog boxes that you constantly have to pop in and out of.
Steve
--
1- The 6502 architecture and instruction set is the only assembly/machine
language environment I can claim to be conversant in.
2- WYSIWYG in the classic retail-sales sense, not in the desktop publishing
sense.
3- LDAO is guilty of this, I think. In the VEC, you can't select a part,
so you can't access the functions of the context menu except by
right-clicking the mouse on the part.
4- Power Designer and DSSArchitect are two that spring to mind.
|
|
Message has 3 Replies:
 | | Re: Perl rules!
|
| (...) Perhaps you're right, but that jump instruction is relative to the Instruction Pointer which uses a full register. The segmented way you are relative to a segment which only uses half a register. You can take advantage of this half register (...) (27 years ago, 22-Jul-99, to lugnet.off-topic.geek)
| | | Re: Perl rules!
|
| (...) OK, at this point I'll have to take your word for it. The segmented address still seems like a high price to pay for a half-register. (...) Maybe this is the case in LEdit -- I don't have knowledge of that source code. But in LDraw, the (...) (27 years ago, 22-Jul-99, to lugnet.off-topic.geek)
| | | Re: Perl rules!
|
| (...) That's a good suggestion. Especially because it's something I can make LDAO do. As opposed to some really nifty suggestions I've received which would require changing LDLite, or would be so processing-intensive that they aren't practical. (...) (27 years ago, 22-Jul-99, to lugnet.off-topic.geek, lugnet.cad.dev)
|
Message is in Reply To:
| | Re: Perl rules!
|
| (...) Relative pointers in the same segment take up half as much space as a 32 bit pointer. Depending on what you're doing, the savings here can be considerable. .COM programs used nothing but relative pointers and fit in less than 1 64K segment. (...) (27 years ago, 21-Jul-99, to lugnet.off-topic.geek)
|
433 Messages in This Thread:
(Inline display suppressed due to large size. Click Dots below to view.)
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
This Message and its Replies on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
