Open Letters #04 21 октября 2000

Nemo open letters № 4.7

   Hello respected firm Nemo!
Wrote you a programmer with experience.
 1). I would like to book you on the PRICE LIST
hardware and software.
 2). Please send me a description GENERAC SOUND, and GENERAL 
SOUND II, if any. 

   I myself have seen many models SPECTRUM
compatible computers, and finally purchased and your KAY-1024. 
I got a definite opinion about the characteristics this clone. 
Well for starters are good at I think part KAY.

 1. 1. MB-RAM is cool, but it would be better
supply and MW (or 8 Mb) ((probably greed)):
 2. Very well-tuned and synchronized TURBO REGIME 7.0. MHz.
 3. TURBO floppy disk controller
 4. Fully implemented LPT port.
 5. Three slots at the periphery (but better
more: 5 -10).
 6. A well-balanced video
(Very handy!).
 7. Operated only from the 5 V B / P or batteries
(COOL).
 8. Indications for treatment and BETA-TURBO
TURBO mode (convenient).
 9. RAM DISK (C: /) (a necessary thing).
 10. I ask BOOT HDD and LPT-PRINTER in BASIC 128 (needed in our 
time). 

   Well, now about the shortcomings that I
discovered during the operation of your
KAY'ya and suggestions on each item. I
I hope you listen to the opinion of the lead programmer in 
Bratsk. (Sorry for immodesty, but I also electronics, 
circuitry, self-taught). 

 1. This applies to the ports configuration
# 1FFD and # 7FFD and system # FD. The point
that when direct appeal to the port # FD
LD A, # 11 OUT (# FD), A recording is
to port # 1FFD, but should the port # 7 FFD.
Because of that 30-45% of the programs does not work
at all. I see three options for removing
this deficiency:

 1). Software Solution:
 1. short addressing replaced by full
# FD replaced by # FFD:
 2. Commands like LD A, # 11 OUT (# FD), A
LD A, # 51 OUT (# FD), A., but this method is a global 
shortage: the fact that it is very much divorced component type 
PENTAGON PENTAGON 512 or 1024, which use 5, 6, 7 - bits under 
the extended memory. 

 2). Hardware replacement address additional port configuration 
(# 1FFD) although would # DFFD, but rather on or # 1FFA

something like that.

 3). Make hardware shutdown treatment to an additional port 
(which I did) or make hardware and software

blocking the port # 1FFD as made port
# 7FFD BIT DS.
 2. Another drawback is
very uncomfortable managing extended
Memory:
(Port # 7FFD BIT D7
Port # 1FFD BIT DS - change) error. see
below.
Port # 7FFD page RAM in this

             D0! D1! D2! D3! D4! D5! D6! D7!
Port # 7FFD page RAM! screen! 128 /! BL! CBA! QG /!

            in the 0-th sector! ! TRDOS! ! ! !

     # 1FFD RAM /! Q8 /! TURBO! ROMS! BANK! A17! A18! STROBE!

             ROM! ! MODE! ! RAM! ! ! !
 3. Error circuit designers in the keyboard port # XX FE 
reading BIT DS should be at 1 instead of 0 because it does not 
work somewhere 30% of the programs.  4. When reading the port # 
FE IN A, (# FE) should be read in all the series keyboard.

Somewhere 40% of the programs click on the check
any key that way.
 5. When reading from the port # BFFD read
# 00 instead of # FF, as a result of some
Program concluded that there is no
AV-3-8910 (12), with corresponding consequences.
 6. Controller BETA-TURBO immediately takes one slot (not 
profitable). I wish you propose to make the controller 
BETA-TURBO multicards composed of includes (and it takes only 
one slot):  1). BETA-TURBO-HD (drive with volume 1yu44

CF and CF 1b2).
 2). Controller HDD (MASER / SLAVE) on 4-D
HDD or CD-ROM.
 3). AT/HT/PS/2 controller - keyboards.
 4). At least 2-com port 1 and the second.
 5) ISA slot or 2-I ISA slots for IBM
devices (eg modem USR SPORSTER).

   I even would have bought this for multicards
additional cost. COM 1 would work
KEM8STON MOUSE, and COM 2 as the AY-MOUSE. It would be very 
convenient for ordinary USER'a.  7. TURBO desirable to increase 
to 10,5 MHz or 14 MHz. I, for example, 280 H

Held at the PROFI PLUS 21 MHz COOL.
 8. Length INT does not meet the desired 20 - 32 CYCLES and INT 
is between PENTAGON'ovskim and brand that

particularly badly affected as
view DEMO'K. It would be desirable to tighten INT 
PENTAGON'ovskomu or, at worst, to corporate.  9. 69,000 cycles 
per INT - this is not enough (Roughly). It is desirable to 
catch PENTAGON - it is about 72,000 cycles.

 10. Check the data bus at the time of treatment to the TR-DOS. 
It is not stable, that very bad.

 11. Desirable to expand the menu of the BASIC
128:
 TR-DOS 1024
 TR-DOS 128
 TR-DOS 48
 BASIC 1024 + CALCULATOR
 BASIC 128
 BASIC 48
 BODT HDD and something more useful. Such as CMOS. And I have a 
question for you about button MAGIK: Could you

flash any service as a monitor
on SCORPIO'ne or at least write as
move from ROM TR-DOS (when you click on MAGIK) in ROM 
EXPANSION: ROMS = 1 and 128 ROM ??


  How much do I reflash my ROM. You would
only knew. And I developed a scheme
ROM programming and erasing any UV
capacitance that connects to the LPT-KAY,
but data on the ROM with a capacity greater than 64 kV
I have not, so software is not adapted
a super large chip, if you
have this information, I will be glad to see it through and 
send to you on distribution (by the way, the programmer does not

must be connected to the LPT,
You can stick in the slot).

  Well, actually, all I wanted you
say. You may want to print
my letter in an open letter, I am against
I will not, and will only be happy if someone
support me and wants me to correspond and exchange SOFT'om. 
Sorry for the error, if any. 

                                       N.

665708 04.08.2000


              Dear N!

1. In my opinion it would be reasonable not
"... An experienced programmer, a coder with
experience. Programmers create software
Products ...

2. No software is profiled products
(Goods) - So in general price list only a few positions. This 
companion goods.


3. The attached package includes advertising
and leaf-passport in StS.

4. 8 MB of RAM - it's not greed, and the action
stereotype that the steeper the better. This
not always the case - and sometimes vice versa.
Growth performance does not automatically improve the consumer 
qualities. How do you imagine working with 8 - yes, even with 2 
MB of RAM? To download a 2 MB RAM needed 3

diskettes, and TRDOS not support such a volume - and, 
consequently, it is necessary OS'6 because when such volumes of 
information work differently, rather than with 
strukturamifaylov is impossible. And if there OS'6, then 
required hard drive, and if there is the hard drive, the RAM 
and requires no more than the CTA Twenty-eight Kilobyte. The 
speed of the OS'i - (note: OS'i! - And not

hard disk or RAM) - virtually the same as with RAM, and with a 
screw - to be may, with the exception of some exotic 
applications / tasks. In KAY-1024 is not sold 1 MB RAM, 
kvazidiskovod implemented on 1 MB of RAM. By itself, 1 MB RAM 
in Spectrum'e does not make sense. This PC shny stereotype. 
Need to read less PC shnoy press and often think of my head.

In addition, it is perhaps the only
way of joining with OS'i TRDOS-disc,
which can operate OS'6 TRDOS
disk (disc). All operations where the possibility of OS'i FILE 
(disc) - except for its launch on performance (transfer 
control). When transferring control OS'6 ignored. 

5. And why not a hundred slots? - Namely, 5 -
10? Why does a man two legs, not three
and not four? Why progress is on
place and a person does not grow legs,
number is greater than that available? Why
car always four-wheel drive - and it is personal, individual 
transport, while the truck is ALWAYS more than four wheels - 
and it is not Individual transport? PC computer

stands for "personal computer", but it is neither personal nor
personal computer. Respect
Port # FD - this is nonsense PENTAGON'ovsky
(Or folly, in translation from English.). With this
squandering of the address space
bus, for example, makes no sense - no matter how much
there was no slots. In practice, any
expansion PENTAGON'a anyway -
give rise to multiple toggle switches as architecture is 
CONFLICT and mutually contradictory - that is, no single 
architecture. This IMPLICATIONS DOURI (now in Russian). 
Therefore, the ONLY SOFTWARE elimination. By the way, I might 
add, at the PENTAGON OR LESS Unable to install the hard drive, 
and therefore OS'6 can function only in a very flawed way. 
Therefore, the PENTAGON and used for DEMO Mayking, as anything 
else to make it difficult. Hardware improvements port # FD the 
desired effect will not result because but dope with # FD, in a 
PENTAGON'e still a number of silly things, so software 
compatibility increased only slightly. In addition, there are 
more weighty consideration. PENTAGON'y now

not available - and, in fact, it's antique. Sooner or later 
they will go out operation. Moreover, the faster

more "improvements". Operation
these components is big money, so
mass they can not be - it is exclusive. Therefore, once taken 
into account PENTAGON'ovskuyu nonsense is futile. In St. 
Petersburg they have almost gone - all in the trash. However, 
waiting and other PENTAGON'y in the near future - regardless of 
whether I am to them and what I do, so

as independently and how to them
are their masters and that they are doing.
"What I've done" - ie, set the switch - as you can see, 
PENTAGON'ovskaya nonsense contagious. Congratulations to the 
end of the warranty. 

6. "Managing the extended memory." Extended memory control is 
not necessary, it must be used. It is properly managed

either OS or TRDOS, - as an electron
drive. Direct encoding programs
1 MB of space is meaningless so
that borders on idiocy. In addition,
"Hard" peg to the equipment, which
inevitable when writing programs under the "iron" (and not the 
architecture!) - because made a direct appeal to the hardware - 
making it impossible to further develop the equipment. All 
Programs (or almost all) are written so that when slightest 
development of equipment (which Always associated with some 
changes architecture) will fly into the ditch. You do

do you want? The question itself suggests
the fact that you never write programs,
BUT ONLY CODE. If you have ever
create at least one program, then himself
the question of the uses of 1 MB would seem absurd to you.

7. "Failed to circuit designers. It's not a bug
circuitry. It is deliberately and consciously do a thing. First 
you need to understand that: 

7.1. There are proprietary software, Nagle requiring 0 in this 
bit. Ie on value of this bit No restrictions even

in the corporate prototype. Moreover, in some machines, this 
bit is enabled. 

7.2. Port as external to
computer butt implies ARBITRARY, NEPREDSKAZYEMYE original value.
That's why he and the port. Constants from the port of only 
reads idiots. If you look Professionally made program,

find that the unused bits are masked ALWAYS. It is clear that 
lamer, fucking on the keyboard, can not know - so appearing as 
a result of the "program" is not a benchmark and weakly 
consistent with what is understood under the program. Because 
of these iliotov (and this is an understatement) cripple the 
machine. This consequence of the fact that one bad head

MANY hands rest does not. Since damage to the head causes 
damage to the machine. 

7.3. In the future, perhaps this bit
ports - as well as other untapped -
tozhet be used for its intended purpose, ie, there will be no 
cap ("0" or "1"), and the signal. Therefore, reading

constants "0" there just is defective, as
and reading "1." The problem is that they write lamer - 
corresponding to the result. It's not a reel, just a dolt 
sitting in the cockpit. Here he wrote komandochku - and checked 
- yeah, go. And the fact that it in half an hour will fail in 
its same computer - it never occurred. Because

that he wrote under the iron, rather than the architecture. As 
for the 30% compatibility - a tip: Do not use lamerskim 
shkodingom - and will not be this percentage. 

8. All series keyboards found in the case, if the high byte 
address bus when reading from the port # FE will be all 0.


9. If the program has been tested in the presence of AY
KAY'e, - it is usually found (see,
For example, the distribution copyist XLC).
If it does not find him, then this defect
programmer to write it. (Again
question - a programmer to write?). It is impossible to 
consider all LAMERSKIE Duri, - and no need - why cultivate 
stupidity?. 

10. About 6.1 - 6.5 with no comments
- Otherwise it will hurt. The only thing I will note
that there are real limitations on
existence of which you seem to not even
guess. You are therefore not recommended to discuss this topic.

11. Physical clock speed is not always
reflect the actual performance. Public
have not yet learned to use even 7
MHz. Yes, and how she learns, if 90% of sitting on the junk, in 
which 3,5 MHz and works with the muddle. If the fleet in July 
MHz amounted to at least 50% - this one would think. It is 
useless to do major repairs in the apartment, which

located in a house which a hundred years without major repairs.

10. Duration INT'a - 2 mash team
(EI) or 3 mash. teams (DI). Ie INT
(Its length) is measured in machine
teams, but not in makrosekundah (or cycles). Therefore, its 
duration is the same in TURBO / NORMAL, which is very 
important. Ie such a way to implement INT'a can be regarded as 
specific TURBO-machines - otherwise the software will fail. In 
addition INT synchronized with the CPU. Ie he

always, everywhere and in all KAY'yah same,
which is not the shaper
INT'a on the RC-circuits. In addition, it seems, you
confused the concept of tact and cycle. As for DEMO, it works 
very tightly tied to hardware. Therefore, the software 
compatibility between the clones is always small - this is 
normal, otherwise simply can not be. Demo account

specifics of the video processor, and specificity
GPU - is its deviation from the
Video standard TV. PENTAGON most
"Curve" in this sense, a clone.

11. "... Catch PENTAGON ..." impossible
as this requires disfigure architecture - ie, make it through 
the ass, as is done in PENTAGON'e. In PENTAGON'e 320 lines 
(must be at the standard 312/313), and so there followed INT'y 
not in 50 mS, and by 640/625 = 1.024 x 50 mS = 51,2 mS, which 
is a flagrant contradiction is no longer with telestandartom, 
but with SI system. Seemingly small thing, however:

 1). AY processor will distort the music -
IT IS FATAL, a diagnosis, a name which - PENTAGON.
 2). Connecting the video encoder PAL / SECAM
if possible, with a fair amount challenges - and not on all TVs.
 3). Real Time Program, geared for the INT, will work in an 
unreal time. For example, to implement the clock in the program 
will require software "Patch" to this folly place in the form

correction 3 / 125, because fraction is irreducible,
then every 125 frames (and it is 50 mS x 125 =
6.25 sec.) Correction must be done
0.15 sec. - That, in some cases, unacceptable. Ie the clock 
will "twitch" every second on the seventh one-seventh of a 
second. Now: (69830/69840) x 1.024 = 71506/71506 - it's rude. 
For a more precise definition is necessary to consider the 
error most KAY'ya. The complete code for 312 x KAY'ya

2 = 624, ie, Error 1 = 312 +313 625 625
(71506/71516) h645/624 = 71630/71620 - is
exactly. That's how much is these very
ticks in PENTAGON'e. Moral here is that the speed of the 
machines Equally, - But different ways of measuring them - and

Specifically, the intervals between INT'ami. So what
We catch up?

12. What do you mean by "instability
Data bus?. your intention is similar to
weather forecast or astrological
prognosis. If you mean that when accessing TRDOS'u in one cycle 
TWICE change the data - so it's all Spectrum-machine case. This 
feature of the entrance to the TRDOS. At the time of entrance 
"is broken" machine cycle, and these changes directly

before reading the first byte of the command
ROM TRDOS'a. If the machine is in turbo mode and
ROM access time is more than 150 nS,
are possible glitches. You probably do with them and
faced when copied ROM -
access time was higher - for example, 200 nS. Could be done here
all in sync - but in my opinion, it is
would lead to an unjustified uslovzhneniyu
scheme. Simply put the "fast" ROM -
benefit them in abundance - so no need to
fuss.

13. I do not think that we should expand the menu.
First of all, on the basis of general considerations. All of 
these options you do not need If the program is written 
correctly and is optional HDD boot. But already in the HDD boot 
option - Ie at the entrance to the operating system

- The menu has the size of the directory. "Another
something useful ... "should be there
and not in the Start Menu.

14. With regard to the service monitor. The problem is the 
concept of shadow software. Those versions of concepts that 
were - did not fit. In fact, the main purpose of the existing 
shadow security - is hacking. I would like something more 
positive, constructive. Ie software is not stealing - and 
creating programs. ie shadowing software must be different. A 
this is the ideology of systems - for coders thing 
intellectually disabled. There needs a programmer. 

15. The problem of creating programmer
Again, no circuitry and software. And it needs a program like
Product, not code. It should be: the base of the crystals, the 
base rom, advanced system HELP, introductory and general 
information on ROM - because many do not understand what it is. 
All this can be done only under OS'6yu because it should

be a single medium. Additional information I have, but it is 
unlikely you will happy. The fact that the handling of firmware 
over large circuits "IMPOSSIBLE without OS'i - because a flash 
over TRDOS-size disc. Ie it is NOT operable in TRDOS'e. 
Imagine, for example, a comparison operation of the two portal 
for 27,080 crystal (a 1 MB ROM). The programmer does not have 
to stick in the slot, because for This requires razvirat 
computer. For "Sarah", which are now quite common, it's risky - 
and sometimes fatal.


                                 (C) Nemo

P.S. All that is discussed here repeatedly mentioned in 
previous letters. If you read carefully with "Op. Let.", then 
many questions, probably would disappear by themselves.