Some years ago, the Hampshire County Library service in the UK had two different styles of library tickets. One type was the standard cardboard wallet into which went a ticket identifying the book. On this carboard wallet, about 1.5" square, was the borrowers name and address. When you returned the book, you got the cardboard wallet back, leaving no trace of who borrowed what, and when. The other type of ticket was the 'Fiction Token'. This was a simple, mass-produced plastic card, identical to every other plastic card, which was simply exchanged for fiction titles. You take a book, you give 'em a token. You return the book, they give you a token back, but not the same one. There's no way to track who has what.
This was all removed in the name of efficiency some years back. The current system uses barcodes in books, and barcoded member cards, tying all books to borrowers present and past. Any librarian can browse through your borrowing history, or the history of a book, almost instantaneously.
So, take a backward step for privacy. Replace your lendng libraries computer system with cardboard wallets. When a book is loaned, you do have the borrowers details, but ONLY while the borrower has the item. This allows you to chase borrowers who have not returned items. Once the item is returned, you lose the association. Simple, private, and virtually idiot-proof too. The system doesn't even need electricity. For low value items, such as paperback books, issue 'fiction tokens'. Borrowers get, say, four tokens, and if they want more, they pay the average cost of a paperback for one. Keep a log of who has how many tokens, but nothing more than that. This will catch abuses, but not provide any tracking. Librarians: You're in the library business, not the espionage business! Do your community a favour, and take a step backwards.
While most americans are probably used to paying sales tax on just about anything, books in the UK are currently zero-rated for purposes of VAT (Value added tax (an oxymoron, because it doesn't add value to anything)) VAT is currently charged at 17.5% of the sale price of a taxable item, although this rate sometimes varies depending on what the item is. (domestic fuel for example is taxed at 5%, and beer at some horrible rate I'd rather not think about). However, books that come with CD's do attract VAT at various rates, depending on the value of the CD. I don't know exactly who decides the value of the CD, but if it contains 22 books, it could conceivably have a very high value to some revenue folk. Still cheaper than buying all the books seperately though. I only bring this up because I discovered that a linux reference book I have here which had a RedHat 6 CD in the back which I've never even used has got me for over a pound of tax. Free software? Nyet.
Many slashdotters will no doubt remember in
the early amiga days, CBM shipped a lot of
systems with 68EC030 CPU's. The 'EC' CPU was
allegedly a plain 68030 which had passed CPU
tests, but failed MMU tests. Consequently, the
chip was unable to utilise virtual memory for one
thing.
However, the demand for 68EC030 CPU's outstripped
supply by quite a large amount, and consequently
Motorola took to stamping '68EC030' on the top
of perfectly good, functional, fully-fledged 68030's
and selling them at a cheaper price.
I had one such beastie on my GVP accellerator card
on my A2000 rev B6. It's still around somewhere.
The European Commission has proposed that a directive be issued that updates member states' laws on email, privacy, and a number of other concerns. Various links from EuroCAUCE here
Moveover, electronic mail for direct marketing purposes other than at the request of a subscriber (so-called 'spam'), will be covered by the same type of protection as exists for faxes. This means that spamming will be prohibited except with respect to subscripers who have indicated that they want to receive unsolicited e-mails for direct marketing purposes.
As legislation goes, this document is remarkably clued-up, and also unusually readable. Everyone move to Europe, quick.
Ok, so it's still in the proposal stage, and won't become a directive until given a second reading by the full EU parliament. If you live in Europe, get onto your MEP's now and ask them to support this directive.
NTL at least offer a free service for automatically blocking calls where the number was withheld (but not calls where the number was just unavailable). The message says something like 'The person you are calling is not currently accepting anonymous calls. Please redial without witholding your number'.
BT offer something similar, although I don't know whether or not they charge for that service.
Some people withold their numbers automatically. If you (or anyone you know) does this, and they can't get through as a result, they need to prefix the dialled number with '1470' which is the code which enables caller-ID for one call only. Sort of like the opposite of 141.
Over on this side of the big pond, the cost of calling mobile telephones
is still orders of magnitude higher than making national or local calls, and
in some instances comparable (or even more expensive, if you use some of
the specialist long-distance carriers) to making international calls.
Example: Calls to mobiles are charged at anywhere from 18.90 to 26.05 pence per minute
including tax during the daytime. Compare this to international rate calls
to the USA at 14.00 and Australia at 22.00 pence per minute. (Source: BT price list). The cost alone is usually sufficient to put people of making calls to mobiles. SMS (text messaging) spam is become more popular though.
The UK Direct Marketing Association, which is the self regulatory
body for direct marketeers in the UK run a scheme called the Telephone Preference Service or 'TPS'. From personal experience, I have found this scheme to be particularly effective against junk callers.
The same bunch also run a postal-mail and fax opt-out service, which is apparently also fairly successful, and an Email service, which I'm somewhat sceptical about (not giving my addressesout anyway, I shouldn't be on their lists). (and most of my spam comes from the USA anyhow)
Under the Telecommunications (Data Protection and Privacy Regulations) 1999 it is an offence for any company engaged in direct marketing to call numbers listed on the TPS. Those who do can be 'shopped' straight away on the TPS site.
One difference I noticed with the USA version of the TPS is that in the USA, you can still be called by local traders, charities and religious groups, AND your number only registers for five years. In the UK, local traders, charities and religious groups are included in the act.
The following was recently posted to demon.announce by Malcolm Muir (although I've snipped it a bit; demon.announce is propogated outside of demon.net; so you can probably read the whole post at Yahoo by now, if you really want..
Whilst we have been aware of this possibility for some time now our networks have been designed to take account of this type of event. Naturally, given the issues with KPNQwest, we have recently confirmed that our network can accommodate the loss of Ebone.
We estimate that about 5% of our traffic is currently routed by Ebone and that is mostly to other European sites.
When or if Ebone is turned off we expect traffic to reroute via other connections in London and Amsterdam.
We believe it is unlikely that there are any other networks connected exclusively to Ebone, so we do not expect any destinations to become unavailable in the event that Ebone closes.
In the event that Ebone close it will take time for traffic patterns across Europe to settle down, however we expect the effect will be limited to some traffic following less than optimal routes and occasional hot spots of congestion.
Unfortunately, 'taken to extremes' is just how some high-ranking people around here like it. (The sort of people that complain about code, but never write it). Another thing I particularly dislike is the insistence of putting { and } around every branch of every if, while, for, etc. This makes the function longer, in terms of lines, (because we also insist on braces lining up and being on their own lines) than it needs to be, and consequently I lose declarations off the top of the screen (although I declare stuff as close to where it's used as possible).
My point about// instead of/* and */ is that the latter is portable when you have to support and update some very old systems with quirky compilers. True story: One enterprising induhvidual here once used #ifdef's around two different style, but otherwise identical comments. Duh!. (Work experience/placement students thrown into the deep end of a critical project where nobody has the time to review their work).
We used to have a very good company coding standard, but it only applied to FORTRAN, and helped us avoid the pitfalls of those platforms where compilers would, for instance, allow identifiers with more than six characters, but only actually used the first six (so WIDGET and WIDGET_LENGTH were the same variable). It also rammed home the importance of not relying on implicit declarations (GOD is real, unless declared integer) and all of the INTEGER*2, and INTEGER*4 length related problems that still bug some braindead C implementations today (where 'int' == 'short' or 'int' = 32bits and 'long' is longer. (I know what the standard says, but I've found that many programmers never read the standard)
I work in a coding shop where our company coding standard insists on some microsoft plot called 'Hungarian Notation'.
For the unitiated, this means that all variable names have to be prefixed with letters that indicate it's type. i for int, f for float, ch for char, etc.
The linux CodingStyle file (in the Documentation directory in every kernel source kit since who knows when) slates hungarian notation thus:
Encoding the type of a function into the name (so-called Hungarian notation) is brain damaged - the compiler knows the types anyway and can check those, and it only confuses the programmer. No wonder MicroSoft makes buggy programs.
If you don't have Linux kernel sources on your machine, you can get a copy of CodingStyle here (from the 2.0 kernels).
As a result of all this mucking about, because I tend to look after most of the dynamic memory, linked lists, and low level bit and byte-bashing operations, I end up with variable names with more prefix letters than letters in the name. I really detest this coding standard (which for some reason also forbids the underscore character on the grounds that it looks like a minus. Do you get foo_bar and foo - bar confused? I don't.
I don't agree with all parts of the Linux CodingStyle, especially the bit about brace placement, but it's a good starting point for any C coding standard. Unfortunately, ours was designed by microsoft-centric folk who think that// is the start of a comment.
Ayhow, back to topic, the Linux CodingStyle also contains the distilled wisdom:
Comments are good, but there is also a danger of over-commenting. NEVER try to explain HOW your code works in a comment: it's much better to write the code so that the _working_ is obvious, and it's a waste of time to explain badly written code.
Atleast you vote, which I did too. I live in Hilsea, Portsmouth, which was won by the conservative (for the benefit of non-uk readers, right-wing) party on the strength of two votes. (neither of which were mine) I've never missed an election, but with four out of five people not even bothering to vote, we have a problem.
I'm not saying I agree with any of the provisions of the library filters bill. Personally, I'd far rather see no filters, and have parents surf *with* their kids. (Perhaps they will both learn something?). Of course, there are those who would rather damn something than open their minds a little. Unfortunately, the people that have a point to make and are willing to make a loud noise about it, are the people who end up in public office. Common sense is, from an electoral point of view, rather boring. Sex is, after all, just a (rather enjoyable) part of biology.
I must admit I am a little mystified about the process of lawmaking in the USA. My main point of confusion is the way representatives attach 'riders' to bills before the houses, in such a way as to bring about either the opposite of what is originally intended, or bring into law a ruling on a completely unrelated topic. The Online Personal Privacy Act might mysteriously become the Online Personal Privacy, Coffee minimum caffeine content, and parking ticket fine enforcement act. It's a wonder that you guys ever get any sensible legislation through when you have to search through virtually all legislation out there to find potential threats to your privacy.
Take for instance, this bill (Sorry, this is a remarkably long page) which is allegedly on Banruptcy reform, but which has attached to it, the Methamphetamine Antiproliferation Act of 2000, which to the average non-lawyer european seems an interesting choice of content in a financial bill.
Not to mention the farce surrounding the mandatory installation of internet content filters in schools and libraries attached to a funding bill old news here. I'm sure you can find better examples.
I have not tested other AV products yet, but the code from the securityfocus.com page is detected as 'Exploit-CodeBase' by McAfee VirusScan 4.5.1, engine 4.1.60, DAT 4.0.4196 [10-Apr-2002], and probably earlier versions.
A few weeks ago, I got some spam that made it through my various lameness filters into the general inbox. This irks me enough usually to track down and destroy the sender's account.
To my suprise, it was relayed through a server in doubleclick's canadian office network.
Figuring that doubleclick being a company with a network to protect, would not have sent the spam deliberately, I contacted their admin and informed them of what initially appeared to be a potential open relay.
Some snotty marketroid wrote back and said that if I didn't want the email, I should follow the 'remove link' in the message footer. (Yeah, Like that ever works). I replied, patiently explaining reasons for not responding directly to spam, nor following links from it. (Heck, I don't even load images and other objects from HTML spam, partly for security reasons, but mainly to stop leaking 'tracking' information that confirms that a spam has arrived (not that many do, having been killed by the aforementioned lameness filters). The link was, as usual, utterly useless, as the sender's account had already been terminated by their ISP. (Dang! someone beat me to it)
This told me all I needed to know. Doubleclick support spam, period. Into the blackhole they go...
Well, I said it was netscape-unfriendly. For those of you getting badly-formatted microsoft error pages, try the 'British Coinage' and 'News' links in the left hand frame on the Royal Mint site.
The Royal Mint [Danger: Netscape-unfriendly site] has commemorated Marconi's 100th anniversary in UK coinage with the current issue of the UK Two Pound Coin, News item here and picture of the thing on the right hand side of the first picture on this page
What sort of test is 3300 users? Have they all arranged for them to wander into one area at the same time and all try and connect at once? The system is in dire need of a 3G slashdot effect.
What essentially happens is that four bits are converted into five bits using a small table, in such as way as it is impossible to have more than eight one-bits in a row, or more than two 0 bits in a row.
A 1 bit is then recorded as a phase change, and a zero bit is an absence of phase change.
Split your data into four-bit lumps, look them up in the table, and write the five-bit group codes out end-to-end. You'll never get more than two 0's or more than eight 1's in a row.
Because the recording density on a given medium is limited in terms of the number of phase changes over a given length, you improve over FM (which puts in two phase changes for a 1, and one for a 0) by using GCR, because you don't need so many phase changes, so you can use less area.
Yes, but that's on Double Density media. 2.88 is ED media (one step up from HD). Add in the higher spec media into the equation and you've probably got 4Mb+
> I believe the transfer rate was somewhere around 2 bytes per second as shipped
Not quite, my guess was about 2 seconds for 254 byte sectors (bearing in mind that the pointer to the next track and sector was in the first two bytes of PRG and SEQ files). Through software alone I managed to get the serial bus up to about one track of 21 sectors in 1.5 secons, (roughly 8k per second), but only with the video generation turned off to make memory access timings more reliable. It even got down to 'how many NOP instructions do I need' at each end to synchronise stuff.
As for changing the device number, do you remember the 'official' way? Boot the system, open a channel to device 8, buffer 15, and send an "M-W" command with some junk I can't remember to flip the device ID to 9, send an 'I' command to reiniitialise, and THEN turn on your second drive. I stuck DIP switches on mine.
If DNA is god's signature, then all we need is his credit card number. And behold, let there be sixteen-way xeon systems! And the users did rejoice, and did feast upon the CPU cycles.
Yes, but that was the fault of the daft serial bus arrangement, which could actually be made a heck of a lot faster with some crafty code (using the onboard CPU on the drive, and carefully written assembler on the '64) to load 48k in, say, 10sec. Infocom adventures were one of the first to use disk acceleration code.
The cassette deck on the 64 was also pitifully slow, for a start it saved everything twice, the second copy being used as a verify pass. All this did was double your chance of an error; I wrote a load of code once to speed that up such that it was faster than the floppy drive. Sad thing was that I didn't have enough money to afford a decent assembler and had to write my own in basic. (C64 basic mainly being PRINT, INPUT, PEEK and POKE and not a lot else I recall). I wrote a basic compiler later, which in some circumstances bettered the petspeed compiler and I still prefer 6502 to 8086 assembley.
The Beeb was a nice machine too, especially since it had a built-in assembler.
Slashdot Mirror
Some years ago, the Hampshire County Library service in the UK had two different styles of library tickets. One type was the standard cardboard wallet into which went a ticket identifying the book. On this carboard wallet, about 1.5" square, was the borrowers name and address. When you returned the book, you got the cardboard wallet back, leaving no trace of who borrowed what, and when.
The other type of ticket was the 'Fiction Token'. This was a simple, mass-produced plastic card, identical to every other plastic card, which was simply exchanged for fiction titles. You take a book, you give 'em a token. You return the book, they give you a token back, but not the same one. There's no way to track who has what.
This was all removed in the name of efficiency some years back. The current system uses barcodes in books, and barcoded member cards, tying all books to borrowers present and past. Any librarian can browse through your borrowing history, or the history of a book, almost instantaneously.
So, take a backward step for privacy. Replace your lendng libraries computer system with cardboard wallets. When a book is loaned, you do have the borrowers details, but ONLY while the borrower has the item. This allows you to chase borrowers who have not returned items. Once the item is returned, you lose the association. Simple, private, and virtually idiot-proof too. The system doesn't even need electricity. For low value items, such as paperback books, issue 'fiction tokens'. Borrowers get, say, four tokens, and if they want more, they pay the average cost of a paperback for one. Keep a log of who has how many tokens, but nothing more than that. This will catch abuses, but not provide any tracking. Librarians: You're in the library business, not the espionage business! Do your community a favour, and take a step backwards.
While most americans are probably used to paying sales tax on just about anything, books in the UK are currently zero-rated for purposes of VAT (Value added tax (an oxymoron, because it doesn't add value to anything)) VAT is currently charged at 17.5% of the sale price of a taxable item, although this rate sometimes varies depending on what the item is. (domestic fuel for example is taxed at 5%, and beer at some horrible rate I'd rather not think about).
However, books that come with CD's do attract VAT at various rates, depending on the value of the CD. I don't know exactly who decides the value of the CD, but if it contains 22 books, it could conceivably have a very high value to some revenue folk. Still cheaper than buying all the books seperately though.
I only bring this up because I discovered that a linux reference book I have here which had a RedHat 6 CD in the back which I've never even used has got me for over a pound of tax. Free software? Nyet.
Another route to what would seem to be the right page is here: http://www.abc.net.au/news/2002/07/item20020730140 728_1.htm
However, the demand for 68EC030 CPU's outstripped supply by quite a large amount, and consequently Motorola took to stamping '68EC030' on the top of perfectly good, functional, fully-fledged 68030's and selling them at a cheaper price. I had one such beastie on my GVP accellerator card on my A2000 rev B6. It's still around somewhere.
Of particular importance is the Proposal for a directive concerning the processing of personal data and the protection of privacy in the electronic communications sector (PDF) which says:
As legislation goes, this document is remarkably clued-up, and also unusually readable. Everyone move to Europe, quick.Ok, so it's still in the proposal stage, and won't become a directive until given a second reading by the full EU parliament. If you live in Europe, get onto your MEP's now and ask them to support this directive.
NTL at least offer a free service for automatically blocking calls where the number was withheld (but not calls where the number was just unavailable). The message says something like 'The person you are calling is not currently accepting anonymous calls. Please redial without witholding your number'.
BT offer something similar, although I don't know whether or not they charge for that service.
Some people withold their numbers automatically. If you (or anyone you know) does this, and they can't get through as a result, they need to prefix the dialled number with '1470' which is the code which enables caller-ID for one call only. Sort of like the opposite of 141.
Jim
Over on this side of the big pond, the cost of calling mobile telephones is still orders of magnitude higher than making national or local calls, and in some instances comparable (or even more expensive, if you use some of the specialist long-distance carriers) to making international calls. Example: Calls to mobiles are charged at anywhere from 18.90 to 26.05 pence per minute including tax during the daytime. Compare this to international rate calls to the USA at 14.00 and Australia at 22.00 pence per minute. (Source: BT price list). The cost alone is usually sufficient to put people of making calls to mobiles. SMS (text messaging) spam is become more popular though.
The UK Direct Marketing Association, which is the self regulatory body for direct marketeers in the UK run a scheme called the Telephone Preference Service or 'TPS'. From personal experience, I have found this scheme to be particularly effective against junk callers.
The same bunch also run a postal-mail and fax opt-out service, which is apparently also fairly successful, and an Email service, which I'm somewhat sceptical about (not giving my addressesout anyway, I shouldn't be on their lists). (and most of my spam comes from the USA anyhow) Under the Telecommunications (Data Protection and Privacy Regulations) 1999 it is an offence for any company engaged in direct marketing to call numbers listed on the TPS. Those who do can be 'shopped' straight away on the TPS site. One difference I noticed with the USA version of the TPS is that in the USA, you can still be called by local traders, charities and religious groups, AND your number only registers for five years. In the UK, local traders, charities and religious groups are included in the act.
Egad. At $1/Gb tax courtesy of crazy french hard drive tax, it'd be far too expensive.
I think that says it all, really.
Just try and load your root-kit onto this machine. Whaddya mean ?OUT OF MEMORY AT LINE 10.
Previously discussed on slashdot back here
Unfortunately, 'taken to extremes' is just how some high-ranking people around here like it. (The sort of people that complain about code, but never write it). Another thing I particularly dislike is the insistence of putting { and } around every branch of every if, while, for, etc.
// instead of /* and */ is that the latter is portable when you have to support and update some very old systems with quirky compilers. True story: One enterprising induhvidual here once used #ifdef's around two different style, but otherwise identical comments. Duh!. (Work experience/placement students thrown into the deep end of a critical project where nobody has the time to review their work).
This makes the function longer, in terms of lines, (because we also insist on braces lining up and being on their own lines) than it needs to be, and consequently I lose declarations off the top of the screen (although I declare stuff as close to where it's used as possible).
My point about
We used to have a very good company coding standard, but it only applied to FORTRAN, and helped us avoid the pitfalls of those platforms where compilers would, for instance, allow identifiers with more than six characters, but only actually used the first six (so WIDGET and WIDGET_LENGTH were the same variable). It also rammed home the importance of not relying on implicit declarations (GOD is real, unless declared integer) and all of the INTEGER*2, and INTEGER*4 length related problems that still bug some braindead C implementations today (where 'int' == 'short' or 'int' = 32bits and 'long' is longer. (I know what the standard says, but I've found that many programmers never read the standard)
For the unitiated, this means that all variable names have to be prefixed with letters that indicate it's type. i for int, f for float, ch for char, etc.
The linux CodingStyle file (in the Documentation directory in every kernel source kit since who knows when) slates hungarian notation thus:
If you don't have Linux kernel sources on your machine, you can get a copy of CodingStyle here (from the 2.0 kernels).
As a result of all this mucking about, because I tend to look after most of the dynamic memory, linked lists, and low level bit and byte-bashing operations, I end up with variable names with more prefix letters than letters in the name. I really detest this coding standard (which for some reason also forbids the underscore character on the grounds that it looks like a minus. Do you get foo_bar and foo - bar confused? I don't.
I don't agree with all parts of the Linux CodingStyle, especially the bit about brace placement, but it's a good starting point for any C coding standard. Unfortunately, ours was designed by microsoft-centric folk who think that
Ayhow, back to topic, the Linux CodingStyle also contains the distilled wisdom:
On a related note, however, the BBC is running a story about how smart tickets will be being introduced on the London Underground. If these can work from a few feet away, how many lamp-posts, bollards, and other street furniture will turn into covert tracking devices?
I'm not saying I agree with any of the provisions of the library filters bill. Personally, I'd far rather see no filters, and have parents surf *with* their kids. (Perhaps they will both learn something?). Of course, there are those who would rather damn something than open their minds a little. Unfortunately, the people that have a point to make and are willing to make a loud noise about it, are the people who end up in public office. Common sense is, from an electoral point of view, rather boring.
Sex is, after all, just a (rather enjoyable) part of biology.
I must admit I am a little mystified about the process of lawmaking in the USA. My main point of confusion is the way representatives attach 'riders' to bills before the houses, in such a way as to bring about either the opposite of what is originally intended, or bring into law a ruling on a completely unrelated topic. The Online Personal Privacy Act might mysteriously become the Online Personal Privacy, Coffee minimum caffeine content, and parking ticket fine enforcement act. It's a wonder that you guys ever get any sensible legislation through when you have to search through virtually all legislation out there to find potential threats to your privacy.
Take for instance, this bill (Sorry, this is a remarkably long page) which is allegedly on Banruptcy reform, but which has attached to it, the Methamphetamine Antiproliferation Act of 2000, which to the average non-lawyer european seems an interesting choice of content in a financial bill.
Not to mention the farce surrounding the mandatory installation of internet content filters in schools and libraries attached to a funding bill old news here. I'm sure you can find better examples.
I have not tested other AV products yet, but the code from the securityfocus.com page is detected as 'Exploit-CodeBase' by McAfee VirusScan 4.5.1, engine 4.1.60, DAT 4.0.4196 [10-Apr-2002], and probably earlier versions.
A few weeks ago, I got some spam that made it through my various lameness filters into the general inbox. This irks me enough usually to track down and destroy the sender's account.
To my suprise, it was relayed through a server in doubleclick's canadian office network.
Figuring that doubleclick being a company with a network to protect, would not have sent the spam deliberately, I contacted their admin and informed them of what initially appeared to be a potential open relay.
Some snotty marketroid wrote back and said that if I didn't want the email, I should follow the 'remove link' in the message footer. (Yeah, Like that ever works). I replied, patiently explaining reasons for not responding directly to spam, nor following links from it. (Heck, I don't even load images and other objects from HTML spam, partly for security reasons, but mainly to stop leaking 'tracking' information that confirms that a spam has arrived (not that many do, having been killed by the aforementioned lameness filters). The link was, as usual, utterly useless, as the sender's account had already been terminated by their ISP. (Dang! someone beat me to it)
This told me all I needed to know. Doubleclick support spam, period. Into the blackhole they go...
Well, I said it was netscape-unfriendly. For those of you getting badly-formatted microsoft error pages, try the 'British Coinage' and 'News' links in the left hand frame on the Royal Mint site.
The Royal Mint [Danger: Netscape-unfriendly site] has commemorated Marconi's 100th anniversary in UK coinage with the current issue of the UK Two Pound Coin, News item here and picture of the thing on the right hand side of the first picture on this page
What sort of test is 3300 users? Have they all arranged for them to wander into one area at the same time and all try and connect at once? The system is in dire need of a 3G slashdot effect.
GCR stands for Group Code Recording.
What essentially happens is that four bits are converted into five bits using a small table, in such as way as it is impossible to have more than eight one-bits in a row, or more than two 0 bits in a row.
A 1 bit is then recorded as a phase change, and a zero bit is an absence of phase change.
A typical 4-5 GCR table might look like this:
0000 01001
0001 01010
0010 01011
0011 01101
0100 01110
0101 01111
0110 10010
0111 10011
1000 10101
1001 10110
1010 10111
1011 11001
1100 11010
1101 11011
1110 11101
1111 11110
Split your data into four-bit lumps, look them up in the table, and write the five-bit group codes out end-to-end. You'll never get more than two 0's or more than eight 1's in a row.
Because the recording density on a given medium is limited in terms of the number of phase changes over a given length, you improve over FM (which puts in two phase changes for a 1, and one for a 0) by using GCR, because you don't need so many phase changes, so you can use less area.
I hope that makes some sort of sense.
Yes, but that's on Double Density media. 2.88 is ED media (one step up from HD). Add in the higher spec media into the equation and you've probably got 4Mb+
> I believe the transfer rate was somewhere around 2 bytes per second as shipped
Not quite, my guess was about 2 seconds for 254 byte sectors (bearing in mind that the pointer to the next track and sector was in the first two bytes of PRG and SEQ files). Through software alone I managed to get the serial bus up to about one track of 21 sectors in 1.5 secons, (roughly 8k per second), but only with the video generation turned off to make memory access timings more reliable. It even got down to 'how many NOP instructions do I need' at each end to synchronise stuff.
As for changing the device number, do you remember the 'official' way? Boot the system, open a channel to device 8, buffer 15, and send an "M-W" command with some junk I can't remember to flip the device ID to 9, send an 'I' command to reiniitialise, and THEN turn on your second drive. I stuck DIP switches on mine.
If DNA is god's signature, then all we need is his credit card number. And behold, let there be sixteen-way xeon systems! And the users did rejoice, and did feast upon the CPU cycles.
Yes, but that was the fault of the daft serial bus arrangement, which could actually be made a heck of a lot faster with some crafty code (using the onboard CPU on the drive, and carefully written assembler on the '64) to load 48k in, say, 10sec. Infocom adventures were one of the first to use disk acceleration code.
The cassette deck on the 64 was also pitifully slow, for a start it saved everything twice, the second copy being used as a verify pass. All this did was double your chance of an error; I wrote a load of code once to speed that up such that it was faster than the floppy drive. Sad thing was that I didn't have enough money to afford a decent assembler and had to write my own in basic. (C64 basic mainly being PRINT, INPUT, PEEK and POKE and not a lot else I recall). I wrote a basic compiler later, which in some circumstances bettered the petspeed compiler and I still prefer 6502 to 8086 assembley.
The Beeb was a nice machine too, especially since it had a built-in assembler.