Graphics on PCs

Reinhard Doelz doelz at comp.bioz.unibas.ch
Wed Jul 22 12:58:25 EST 1992


 (The mailer rewrites the address incorrectly... BIONET at EARN.FRCGM51 ... :-) 

In article <1992Jul22.163948.16985 at gserv1.dl.ac.uk>, BIONET at EARN.FRCGM51 writes:
|>  
...
|>  
|>         As regards R. Doelz's reply about molecular graphics:
|>  
|> 1) I don't understand the "eight bits" point. As far as I know, the i486
|>    machines are genuine 32 bits computers ?

I was talking about 8 bit GRAPHICS. If you paint a line in any black 
and white environment, you have 1 bit (on/off == black/white). 

If you use 8 bit graphics, you are usually limited to 256 (2 **8) colors. 
This means that you have no Z-Buffer (for depth sorting) and double-buffer
(for smooth animation). 

You can try to use so-called color tables which enable you to predefine 
colors in a RGB scheme (e.g., 256 from 16 millions). In order to get 
double-buffering you could split the 8 bits in 4 and use the dithering 
technique to get the color intermediates because in this mode you have 
only 4 bit graphics ( resulting in 16 generic colors). 

If you do the z-buffer in software (meaning that you draw objects in front, 
but don't show objects in the rear ) you need a damned fast CPU to calculate
CPK models. You won't look at proteins in CPK but a RIBBON representation
would definitively be a good option for proteins, or Protein/DNA interaction.

To my experience, PCs and Low-end computers hardly deserve the attribute 
'high performance graphics'  if it comes to such challenging problems 
like molecular modelling. I definitively like to see molecules as I 
imagine them to look like. This is not feasible on PCs. 


|> 2) The price is not as high as stated. We've bought last week the following
|>    clone:
|>         * i486-DX2 processor @ 50MHz with coprocessor
|>         * 8 Mb RAM
|>         * 5"1/4 and 3"1/2 floppy units
|>         * 250 Mb hard disk
|>         * 17" high resolution color monitor (1280x1024)
|>         * MS-Dos 5, MS-Windows 3.1 and mouse
|>  
|>         for 27,000 FF (approximately $ 5,000)
|>  

This is cheap in its literal sense. What monitor quality do you get 
for such a price? What keyboard quality? Fan noise? Hard Disk access time? 

I would like to point out that the 250 MB hard disk is too small for SCO UNIX 
as you wanted it in the initial posting. Further, you didn't state the backup 
device (streamer)

 I think it is terrific that you feel happy with this setup, but 
then the question you asked is to compare apples with eggs. 

|>         Here in France, we need at least three times this amount to get the
|> basic minimal configuration of an SGI Indigo ... which, admittedly, is not
|> a PC and has tremendous graphics capabilities if not basic and minimal.
|>  

I did not want to sell you an Indigo. I am in no way affilated with SGI and 
all the words are mine etc etc etc... 

I just wanted to WARN you that, depending on what you want to do, it 
might very fast be a  underpowered system you are looking at. I did my 
first steps in this direction as well. Depending on what you want to do, 
even Tektronix graphics on a VAX can serve your needs. (E.g., Macromodel
jumps into mind here). If you are in a center for genetics, where people 
'just' want to look, good graphics are essential because the human perception 
for molecular models is quite limited if they are presented insufficiently.

Just keep in mind that buying hardware first, and looking for an application 
afterwards, is a very risky route to take. You should have opted for 
SOLUTIONS rather than for a given hardware, and then take your pick after 
having demo'ed  them to all  of your customers. Don't forget to mention 
software maintenance, support and documentation to your collegues. 
I have seen people sitting at PCs for a whole easter holiday as they tried 
to figure out to draw a DNA molecule on the screen unless they figured 
out that the program had hard-coded peptide connectivity only to display 
proteins. 


Good luck, anyway. 

Regards 
Reinhard 

(I sent this by email with the correct address but it bounced). 
-- 
+----------------------------------+-------------------------------------+
|    Dr. Reinhard Doelz            | RFC     doelz at urz.unibas.ch         |
|      Biocomputing                | DECNET  20579::48130::doelz         |
|Biozentrum der Universitaet       | X25     022846211142036::doelz      |
|   Klingelbergstrasse 70          | FAX     x41 61 261- 6760 or 267- 2078     
|     CH 4056 Basel                | TEL     x41 61 267- 2076 or 2247    |   
+------------- bioftp.unibas.ch is the SWISS EMBnet node ----------------+
               -----------------------------------------




More information about the Bio-soft mailing list