Tim Clark

In article <1195725487.9895.0@proxy01.news.clara.net>,
The Natural Philosopher <a@b.c> writes:
> dave @ stejonda wrote:
>> In message <1195674905.17263.0@proxy00.news.clara.net>, The Natural
>> Philosopher <a@b.c> writes
>>> dave @ stejonda wrote:
>>>> In message <1195651491.55997.0@despina.uk.clara.net>, The Natural
>>>> Philosopher <a@b.c> writes
>>>>>> 50KBps is roughly 4Mbps
>>>>>
>>>>> No, its 400KBPS. Exactly.
>>>> always?
>>>>
>>> Always. 8 bits to a byte
>>>
>>> Ok when you start getting into layered protocols over packets over
>>> packets all with CRC and header and routing..that's a different
>>> matter, but BUT he wasn't saying that - bits per second to bytes per
>>> second is 8:1. By definition.
>>
>> When I were growing up that *wasn't* a given. Obviously it is nowadays.
>>
> well..you mean serial protocol etc..?
> baud/10 ~= bytes/sec?

It is true that for most common transmission speeds from 300 to 9600
bps, one divided by 10 to get characters per second, because of the
start bit and stop bit. 110 bps, typically used for teletypes was an
exception - they need an extra stop bit to recover from the trauma of
printing a character. So 110 bps usually meant 10 characters per second.

However, computers themselves used different byte lengths. Although one
more often talked about the word length, byte length varied from one
computer to another. I worked on one where the word length was 24 bits,
holding 4 6-bit bytes (more usually called characters). Another with a
word length of 48 bits could treat it as either 8 6-bit bytes or 6 8-bit
bytes. If my memory recalls correctly, I believe there was a computer
with 36 bit words, often divided into 4 9-bit bytes.

Early standards often used the term "octet" to mean unambiguously 8
bits due to the variation in byte length.

--
Tim Clark

dave @ stejonda

In message <d4fh15-kb4.ln1@mint.tcict.co.uk>, Tim Clark
<Timothy.Clark@btinternet.com> writes
>In article <1195725487.9895.0@proxy01.news.clara.net>,
> The Natural Philosopher <a@b.c> writes:
>> dave @ stejonda wrote:
>>> In message <1195674905.17263.0@proxy00.news.clara.net>, The Natural
>>> Philosopher <a@b.c> writes
>>>> dave @ stejonda wrote:
>>>>> In message <1195651491.55997.0@despina.uk.clara.net>, The Natural
>>>>> Philosopher <a@b.c> writes
>>>>>>> 50KBps is roughly 4Mbps
>>>>>>
>>>>>> No, its 400KBPS. Exactly.
>>>>> always?
>>>>>
>>>> Always. 8 bits to a byte
>>>>
>>>> Ok when you start getting into layered protocols over packets over
>>>> packets all with CRC and header and routing..that's a different
>>>> matter, but BUT he wasn't saying that - bits per second to bytes per
>>>> second is 8:1. By definition.
>>>
>>> When I were growing up that *wasn't* a given. Obviously it is nowadays.
>>>
>> well..you mean serial protocol etc..?
>> baud/10 ~= bytes/sec?
>
>It is true that for most common transmission speeds from 300 to 9600
>bps, one divided by 10 to get characters per second, because of the
>start bit and stop bit. 110 bps, typically used for teletypes was an
>exception - they need an extra stop bit to recover from the trauma of
>printing a character. So 110 bps usually meant 10 characters per second.
>
>However, computers themselves used different byte lengths. Although one
>more often talked about the word length, byte length varied from one
>computer to another. I worked on one where the word length was 24 bits,
>holding 4 6-bit bytes (more usually called characters). Another with a
>word length of 48 bits could treat it as either 8 6-bit bytes or 6 8-bit
>bytes. If my memory recalls correctly, I believe there was a computer
>with 36 bit words, often divided into 4 9-bit bytes.
>
>Early standards often used the term "octet" to mean unambiguously 8
>bits due to the variation in byte length.
>
thanks - whilst I have a vague recollection of all that I couldn't
remember enough detail to respond to TNP.

--
dave @ stejonda

The Natural Philosopher

dave @ stejonda wrote:
> In message <d4fh15-kb4.ln1@mint.tcict.co.uk>, Tim Clark
> <Timothy.Clark@btinternet.com> writes
>> In article <1195725487.9895.0@proxy01.news.clara.net>,
>> The Natural Philosopher <a@b.c> writes:
>>> dave @ stejonda wrote:
>>>> In message <1195674905.17263.0@proxy00.news.clara.net>, The Natural
>>>> Philosopher <a@b.c> writes
>>>>> dave @ stejonda wrote:
>>>>>> In message <1195651491.55997.0@despina.uk.clara.net>, The Natural
>>>>>> Philosopher <a@b.c> writes
>>>>>>>> 50KBps is roughly 4Mbps
>>>>>>>
>>>>>>> No, its 400KBPS. Exactly.
>>>>>> always?
>>>>>>
>>>>> Always. 8 bits to a byte
>>>>>
>>>>> Ok when you start getting into layered protocols over packets over
>>>>> packets all with CRC and header and routing..that's a different
>>>>> matter, but BUT he wasn't saying that - bits per second to bytes per
>>>>> second is 8:1. By definition.
>>>>
>>>> When I were growing up that *wasn't* a given. Obviously it is nowadays.
>>>>
>>> well..you mean serial protocol etc..?
>>> baud/10 ~= bytes/sec?
>>
>> It is true that for most common transmission speeds from 300 to 9600
>> bps, one divided by 10 to get characters per second, because of the
>> start bit and stop bit. 110 bps, typically used for teletypes was an
>> exception - they need an extra stop bit to recover from the trauma of
>> printing a character. So 110 bps usually meant 10 characters per second.
>>
>> However, computers themselves used different byte lengths. Although one
>> more often talked about the word length, byte length varied from one
>> computer to another. I worked on one where the word length was 24 bits,
>> holding 4 6-bit bytes (more usually called characters). Another with a
>> word length of 48 bits could treat it as either 8 6-bit bytes or 6 8-bit
>> bytes. If my memory recalls correctly, I believe there was a computer
>> with 36 bit words, often divided into 4 9-bit bytes.
>>
>> Early standards often used the term "octet" to mean unambiguously 8
>> bits due to the variation in byte length.
>>
> thanks - whilst I have a vague recollection of all that I couldn't
> remember enough detail to respond to TNP.
>

MM. I can remember those times too, but never having non 8 bit combos
(octets) called 'bytes'. Usually '12 bit word' or something like that.

Perhaps you had better start a discussion in Wikipedia, because they
don't mention it either.

Not saying you are wrong, just that I can't find any evidence to
support your claim..

Tim Clark

In article <1195815041.37430.0@iris.uk.clara.net>,
The Natural Philosopher <a@b.c> writes:
>> In message <d4fh15-kb4.ln1@mint.tcict.co.uk>, Tim Clark
>> <Timothy.Clark@btinternet.com> writes
>>> However, computers themselves used different byte lengths. Although one
>>> more often talked about the word length, byte length varied from one
>>> computer to another. I worked on one where the word length was 24 bits,
>>> holding 4 6-bit bytes (more usually called characters). Another with a
>>> word length of 48 bits could treat it as either 8 6-bit bytes or 6 8-bit
>>> bytes. If my memory recalls correctly, I believe there was a computer
>>> with 36 bit words, often divided into 4 9-bit bytes.

> MM. I can remember those times too, but never having non 8 bit combos
> (octets) called 'bytes'. Usually '12 bit word' or something like that.
>
> Perhaps you had better start a discussion in Wikipedia, because they
> don't mention it either.
>
> Not saying you are wrong, just that I can't find any evidence to
> support your claim..

The Wikipedia entry on "Byte" is now seems full of examples of non 8-bit
bytes - I didn't touch it! It mentions 6, 7, 8, 9 and even 12-bit bytes.
http://en.wikipedia.org/wiki/Byte

I agree these entities weren't usually called "bytes" though. In the
case of the first computer I mentioned (the one with 24 bit words
divided into 4 6-bit bytes was an Elliott, later ICL, 4130), the term
"byte" would never be used. These 6-bit units would always be called
"characters".

The curious machine with the 48 bit words, a Burroughs B6700, like
others in its family, had a huge instruction set and hardware capable of
addressing either 6-bit or 8-bit portions of these words. Again it was
more usual to refer to these 6 and 8 bit items as characters most of the
time, but they were definitely called bytes too - especially when
dealing with its data comms.

--
Tim Clark

Pete L

On 24 Nov, 01:10, "Tim Clark" <Timothy.Cl...@btinternet.com> wrote:
> In article <1195815041.3743...@iris.uk.clara.net>,
> The Natural Philosopher <a...@b.c> writes:
>
>
>
> >> In message <d4fh15-kb4....@mint.tcict.co.uk>, Tim Clark
> >> <Timothy.Cl...@btinternet.com> writes
> >>> However, computers themselves used different byte lengths. Although one
> >>> more often talked about the word length, byte length varied from one
> >>> computer to another. I worked on one where the word length was 24 bits,
> >>> holding 4 6-bit bytes (more usually called characters). Another with a
> >>> word length of 48 bits could treat it as either 8 6-bit bytes or 6 8-bit
> >>> bytes. If my memory recalls correctly, I believe there was a computer
> >>> with 36 bit words, often divided into 4 9-bit bytes.
> > MM. I can remember those times too, but never having non 8 bit combos
> > (octets) called 'bytes'. Usually '12 bit word' or something like that.
>
> > Perhaps you had better start a discussion in Wikipedia, because they
> > don't mention it either.
>
> > Not saying you are wrong, just that I can't find any evidence to
> > support your claim..
>
> The Wikipedia entry on "Byte" is now seems full of examples of non 8-bit
> bytes - I didn't touch it! It mentions 6, 7, 8, 9 and even 12-bit bytes.http://en.wikipedia.org/wiki/Byte
>
> I agree these entities weren't usually called "bytes" though. In the
> case of the first computer I mentioned (the one with 24 bit words
> divided into 4 6-bit bytes was an Elliott, later ICL, 4130), the term
> "byte" would never be used. These 6-bit units would always be called
> "characters".
>
> The curious machine with the 48 bit words, a Burroughs B6700, like
> others in its family, had a huge instruction set and hardware capable of
> addressing either 6-bit or 8-bit portions of these words. Again it was
> more usual to refer to these 6 and 8 bit items as characters most of the
> time, but they were definitely called bytes too - especially when
> dealing with its data comms.
>
> --
> Tim Clark

Thanks chaps for all your suggestions and advice. I have finally got
Sky to admit there is a problem on the line. They are investigating.
It sounds as though the problem is that the data coming down the line
is faster than the line can cope with and is causing 'a blockage'.
When the internet is running well I tested 4.7mb whilst apparently the
copper line can only cope with 2Mb as I'm about 3 miles from the
exchange. Presumably they will slow my data feed down a bit and
hopefully I will then get some stability - we shall see!

The Natural Philosopher

Tim Clark wrote:
> In article <1195815041.37430.0@iris.uk.clara.net>,
> The Natural Philosopher <a@b.c> writes:
>>> In message <d4fh15-kb4.ln1@mint.tcict.co.uk>, Tim Clark
>>> <Timothy.Clark@btinternet.com> writes
>>>> However, computers themselves used different byte lengths. Although one
>>>> more often talked about the word length, byte length varied from one
>>>> computer to another. I worked on one where the word length was 24 bits,
>>>> holding 4 6-bit bytes (more usually called characters). Another with a
>>>> word length of 48 bits could treat it as either 8 6-bit bytes or 6 8-bit
>>>> bytes. If my memory recalls correctly, I believe there was a computer
>>>> with 36 bit words, often divided into 4 9-bit bytes.
>
>> MM. I can remember those times too, but never having non 8 bit combos
>> (octets) called 'bytes'. Usually '12 bit word' or something like that.
>>
>> Perhaps you had better start a discussion in Wikipedia, because they
>> don't mention it either.
>>
>> Not saying you are wrong, just that I can't find any evidence to
>> support your claim..
>
> The Wikipedia entry on "Byte" is now seems full of examples of non 8-bit
> bytes - I didn't touch it! It mentions 6, 7, 8, 9 and even 12-bit bytes.
> http://en.wikipedia.org/wiki/Byte
>
> I agree these entities weren't usually called "bytes" though. In the
> case of the first computer I mentioned (the one with 24 bit words
> divided into 4 6-bit bytes was an Elliott, later ICL, 4130), the term
> "byte" would never be used. These 6-bit units would always be called
> "characters".
>
> The curious machine with the 48 bit words, a Burroughs B6700, like
> others in its family, had a huge instruction set and hardware capable of
> addressing either 6-bit or 8-bit portions of these words. Again it was
> more usual to refer to these 6 and 8 bit items as characters most of the
> time, but they were definitely called bytes too - especially when
> dealing with its data comms.
>
Fascinating!

I never heard 'byte' used before the Intel 8008 came along..

harry

On Thu, 22 Nov 2007 23:30:13 GMT, "Tim Clark"
<Timothy.Clark@btinternet.com> wrote:

>In article <1195725487.9895.0@proxy01.news.clara.net>,
> The Natural Philosopher <a@b.c> writes:
>> dave @ stejonda wrote:
>>> In message <1195674905.17263.0@proxy00.news.clara.net>, The Natural
>>> Philosopher <a@b.c> writes
>>>> dave @ stejonda wrote:
>>>>> In message <1195651491.55997.0@despina.uk.clara.net>, The Natural
>>>>> Philosopher <a@b.c> writes
>>>>>>> 50KBps is roughly 4Mbps
>>>>>>
>>>>>> No, its 400KBPS. Exactly.
>>>>> always?
>>>>>
>>>> Always. 8 bits to a byte
>>>>
>>>> Ok when you start getting into layered protocols over packets over
>>>> packets all with CRC and header and routing..that's a different
>>>> matter, but BUT he wasn't saying that - bits per second to bytes per
>>>> second is 8:1. By definition.
>>>
>>> When I were growing up that *wasn't* a given. Obviously it is nowadays.
>>>
>> well..you mean serial protocol etc..?
>> baud/10 ~= bytes/sec?
>
>It is true that for most common transmission speeds from 300 to 9600
>bps, one divided by 10 to get characters per second, because of the
>start bit and stop bit. 110 bps, typically used for teletypes was an
>exception - they need an extra stop bit to recover from the trauma of
>printing a character. So 110 bps usually meant 10 characters per second.
>
>However, computers themselves used different byte lengths. Although one
>more often talked about the word length, byte length varied from one
>computer to another. I worked on one where the word length was 24 bits,
>holding 4 6-bit bytes (more usually called characters). Another with a
>word length of 48 bits could treat it as either 8 6-bit bytes or 6 8-bit
>bytes. If my memory recalls correctly, I believe there was a computer
>with 36 bit words, often divided into 4 9-bit bytes.
>
>Early standards often used the term "octet" to mean unambiguously 8
>bits due to the variation in byte length.


I totally agree - the 24 bits 4*6 wasn't a Philips DS714 Message
Switch was it?

I also remember using the term "8 bit bytes" to avoid any doubt

Graham Murray

harry <me@privacy.net> writes:

> I totally agree - the 24 bits 4*6 wasn't a Philips DS714 Message
> Switch was it?

Or an ICL 1900?

Bob Eager

On Sat, 24 Nov 2007 01:10:14 UTC, "Tim Clark"
<Timothy.Clark@btinternet.com> wrote:

> I agree these entities weren't usually called "bytes" though. In the
> case of the first computer I mentioned (the one with 24 bit words
> divided into 4 6-bit bytes was an Elliott, later ICL, 4130), the term
> "byte" would never be used. These 6-bit units would always be called
> "characters".

I was going to ask if it was a 4130, rather than the more common 1900.
Did a lot of work on 4130s...knew them inside out (too much so for the
system manager at university). Later on I BECAME the system manager...!

> The curious machine with the 48 bit words, a Burroughs B6700, like
> others in its family, had a huge instruction set and hardware capable of
> addressing either 6-bit or 8-bit portions of these words. Again it was
> more usual to refer to these 6 and 8 bit items as characters most of the
> time, but they were definitely called bytes too - especially when
> dealing with its data comms.

And then you have the PDP-10, with variable length 'bytes'....

But early non-IBM stuff very rarely used the term 'byte'...e.g. the
PDP-8 with 12 bit words and 6 bit characters...

--
[ 7'ism - a condition by which the sufferer experiences an inability
to give concise answers, express reasoned argument or opinion.
Usually accompanied by silly noises and gestures - incurable, early
euthanasia recommended. ]

Bob Eager

On Sat, 24 Nov 2007 10:58:09 UTC, The Natural Philosopher <a@b.c> wrote:

> Fascinating!
>
> I never heard 'byte' used before the Intel 8008 came along..

Then you missed out on many years of IBM mainframes, some of PDP-11s,
...

--
[ 7'ism - a condition by which the sufferer experiences an inability
to give concise answers, express reasoned argument or opinion.
Usually accompanied by silly noises and gestures - incurable, early
euthanasia recommended. ]