EXTRA BITS - Printing and Typesetting History - Computerphile

The GSI: A Revolutionary Typing System

The GSI, short for Graphic Systems Inc, was a revolutionary typing system made by a small firm based in Massachusetts. Forgotten to ask Brian exactly how much it cost, we can only speculate that it must have been tens of thousands of dollars. Despite its high price tag, the GSI was capable of producing high-quality output, making it an attractive option for those looking for a reliable typing solution.

Development at Bell Labs

The computer science section at Bell Labs developed an extended text formatting system called trof, which was capable of type-setting to this machine. Although not a dedicated mini-computer in its own right, the GSI could be connected to other systems via parallel port interfaces or used with specialist electronics. This made it possible for users to drive the device directly from other systems, such as the PDP-11.

The advent of trof in the late 1960s and early 1970s marked an important milestone in the development of typing systems. Trof was capable of producing high-quality text output, similar to modern-day typewriters, but with greater flexibility and accuracy. The system used a technique called "enroth," which allowed for precise control over font sizes and styles.

The emergence of trof also led to the development of more advanced typing systems, including TR (typesetter version of rough). TR was capable of producing even higher-quality output than trof, making it an attractive option for those requiring precise text formatting. One notable example is the use of TR in a 1972 University of Nottingham examination cover sheet.

Advancements in Imaging Technology

As the GSI and its successors became more widespread, the need for improved imaging technology became apparent. The third generation of typing systems addressed this issue by using high-resolution cathode-ray tubes (CRTs) to produce images of characters. These CRTs were capable of producing incredibly detailed images, with resolutions as high as 972 dots per inch.

The imaging process involved exposing silver bromide film or photographic negative paper in bands, which were then mechanically moved to create different columns on the page. This process required high precision and accuracy, as even slight misalignments could result in distorted text. To address this issue, specialized gears with no backlash were used to ensure accurate alignment.

The advent of CRT imaging technology marked a significant improvement over earlier typing systems, which relied on simple character matrices or phototypesetting machines. The new technology enabled the creation of high-quality text documents with precision and accuracy, paving the way for modern-day printing and publishing applications.

Legacy and Replacement

By the late 1970s, the GSI had become outdated, and a replacement was needed. In response to this need, Bell Labs developed more advanced typing systems that utilized mini-computers and stored font data in computer files. These newer systems used high-resolution CRT imaging technology to produce precise text output.

The development of these new systems marked an important milestone in the evolution of typing technologies. They paved the way for modern printing and publishing applications, which continue to rely on advanced typography and imaging techniques to produce high-quality documents. The legacy of the GSI can be seen in the many modern typing systems that followed in its footsteps, each building upon the innovations of their predecessors to create more accurate and precise text formatting solutions.

"WEBVTTKind: captionsLanguage: enperhaps it might be a good idea for those of you who really don't know an awful lot about this to just step back a bit and say well how did type setting and printing come about anyway and I'm sure many of you out there know that it goes back hundreds if not thousands of years in the Far East particularly it was possible to uh make marks on paper for Chinese Japanese Arabic characters 100 hundreds and hundreds of years ago those of in Us in Western Europe often fall into the Trap of thinking well really printing was invented in 1456 I think it was by Johan Guttenberg in Ms in Germany and of course he produced the famous Gutenberg Bibles but I think in all fairness you've got to say he didn't invent it he rediscovered it perhaps he refined the techniques and so on so to give you some idea I've got here a sample of wooden type and if you look down here it's actually for a font called caslon which some of you may even have the modern form of under windows or on your Macintosh or whatever that's just a display case but it it shows in many ways the way you used to have to set your type you had to have a wooden frame as big as the page size that you wanted if you had type of this huge size fine you could even handle it with your fingers and put it in in exactly the uh layout and format you wanted most of the time you didn't have the luxury of anything that big that is metal type it's metal type for a font called uni there it's effectively a if you like a sort of French equivalent of the Swiss helvetica type it's sand serif very clean very nice but just look at the size of it this is 10o univ there can you imagine that if you wanted to set one of your your papers or your letters with individual pieces of metal type you would be down to having a wooden frame armed with your Nimble fingers and tweezers maybe if you're going to set at 10 point you'll be picking these up putting them in lines having to accept the fact if you think about it you're going to put paper on top of these and in make an impression therefore these have to be mirror images in order to look right when you take the paper off and typically if you're on a A4 size page you're looking at something 25 times the area of this all to be done by hand and all to be made absolutely perfect no typos no nothing and how do you space the lines well up at the top of this box there's some flat looking pieces in there and those are pieces of what nowadays will be called leading if they're made in lead type which I don't know if this is but anyway the leading is the line spaces that go between your lines of type so you have to insert all those you have to make the lines look even the next stage is to gently roll ink all over it not too much ink you don't want blocks to appear on your paper and you would then put it in a press very much like an apple press or a cider press it's got a great big plate on it it's got one of those screw threads with a big handle at the top and the idea is uniformly and under high pressure to press the paper evenly onto the ink type and then you take the pressure off pull off the paper and if you're skilled enough you've got a beautifully ined version of what's there in your type from guttenberg's time for the next 400 years nothing fundamentally changed what changed in all of this in the late 19th century was that the setting of metal type instead of having to be done by hand became mechanized methods were devised for using these type samples almost as Masters and been able to replicate copies of them in a thing called a typ caster and once you got lots and lots and lots and lots of type enough to do a whole version of The Washington Post or the New York Times or whatever then what you needed was some way of mechanizing the placement of this replica type that machine was invented in Germany in about 1880 I think or thereabouts by a gentleman called Mar mgala he founded a very famous company called mgala the parent company but the name of the machine that did all this was the liner type and as its name implies it set lines of type which is exactly what the newspaper industry wanted a little bit later a rival came along worked under different principles which tended to Corner over the years rather more the if you like quality end of the book Market rather than newspapers that had to be bashed off every single day of the week that was known as the monotype machine I'm not saying there weren't others there were plenty of others because um type was a a good way to make a good living so long as you had the craftsmanship and the skills and so on but the big two were certainly monotype and linotype like the General Motors and Ford there were others but those were the big big names well Apple and Windows yeah apple and Microsoft Maybe so that technology persisted for about another 50 60 70 years uh of just setting metal type and putting in a printing press pressing the paper down and and so on but the next big revolution that came along from about the middle 1940s onwards was the idea of having your type not as metal but as if you like photographic images of the characters so if you can imagine it inside a big light tight box you would have a on a circular drum a strip of film that's got the whole shall we say of times Roman upright at the top and then below it times italic all the characters times bold times bold italic not all worked on exactly this uh principle but it gives you the idea and certainly the one they got at Bell labs this sort of second generation type Setter was very much like this so I'll stick to this description you've got all of your characters in the current font available as photographic images what you've now got to do is to transfer the photographic image of the capital letter s shall we say from the strip onto a piece of photosensitive paper at the back of the box and remember this is all light tight so what you have is an unbelievably complex system of mirrors and lenses capable of either shrinking or expanding the master letter s to be either 36 point or 6 point or whatever you're wanting so the sizing was done optically by magnifying and diminishing and when you were happy that you'd got the letter s to the right size this all takes place in milliseconds I might say with great noise in the background then you say right right I'm ready to print my letter s so you would fire a strobe a strobe light would take the image of the letter s focus it down through a fiber optic bundle and the fiber optic bundle would be positioned just above the photographic paper at the back at exactly the place where the letter s needed to be imaged and then you'd move the bundle a little bit off to the right and then maybe you do a lowercase letter A next door to it so if you can probably imagine it was mechanized but it wasn't Ultra fast you've got a lot of things to move a lot of synchronization to do in firing lights and so on but nevertheless it worked and monotype and linotype had Machines working on this principle but they weren't the only ones and the interesting thing is that at Bell labs in the early 1970s they actually had one of these true phototype Setters it was made by a small firm called graphic Systems Inc of Massachusetts so forgotten to ask Brian exactly how much it cost but is I would reckon tens of thousands of dollars armed with that the computer science section at Bell Labs developed an extended text formatting system called trof which was capable of type setting to this machine I should say that the GSI was just about computed drivable as it was supplied to you the idea was to drive it completely off paper tape which you prepared offline however it was possible to buy some extra electronics it was not a dedicated mini computer even in those days it's far too early it was specialist Electronics you could buy which would enable you to interface a parallel port type interface um off a pdp11 shall we say and to drive it directly from a pdp1 and that's exactly what bell Labs did so there they were with a text format which had been developed in the late 60s early 70s there were a lot of them around there's one at MIT I think called runoff Joo sna at Bells developed one called rough and what happened in those early days was if you ran ordinary rough normal rough as it were just to a text formatter that wasn't a types Setter again and it's back at this sort of quality that was called enroth okay but if in the future you wanted much better quality material you would have something developed called trough and TR standing for types Setter version of rough was capable of producing properly types set output looking just like this that's the way it was at Bell Labs from about 1972 to about 197 79 there was tough and its two pre-processors a guy there called Mike lesque developed one called tuble for typees setting tables you wrote in a tabular language how he wanted your tables um laying out and tubel's job was to translate your highlevel spec into lowlevel t- code and to squirt it into t- Ruffles Brian and his collaborator at the time Linda cherry wrote another one uh which was really just made your jaw drop open to actually typet mathematics and it's hard for me to believe actually even now that with the aid of eqn as of 1972 vintage you could very easily types set a page of that sort of complexity this is actually from a little bit later on it's University of Nottingham examination cover sheet but that the sort of mathematics type setting you could do with using eqn to drive this t- off program by the end of seven or eight years of very very heavy usage the poor old GSI cat was on its knees and uh Brian and his gang decided that they just had to have a replacement and by that stage the third generation had come along the third generation of types Setters actually used mini computers and actually represented the fonts as data inside a computer file the only problem then is well it's all very wonderful but how do you actually get an image how do you make an image how you know what do you do well in the third generation typically you used a very high Precision cathode ray tube to image your characters of course as you all know cathet tubes were the technology for televisions for many many years and yes you could make a catheter a tube you still can to be very high Precision indeed and by high Precision I'm talking about shall we say 972 dots to the inch now if you're going to make it that high Precision you are certainly not going to have a widescreen TV at that sort of resolution not in those days anyway this was where the actual Imaging took place and here there is a cartridge holding the silver bromide uh positive paper or your photographic negative film and underneath this black cover is the high resolution cathet tube somewhere yes is a picture of the highrose cathet which is like a letter box and essentially yeah and essentially what happened was you exposed the bromide paper in bands and and when you finished doing the current band you mechanically moov the bromide away from it if you wanted to do superscript and subscript work just by as it were dodging and weaving within the letter box that could be done without moving the bromide and that was very much favored mechanical alignment problems were a big big issue you had to have very high quality gears with no backlash because if you're setting a two column newspaper paper and if you decide to do it by doing the left hand column first and then winding back the bromide to the top and doing the right hand column next you didn't half uh come across alignment problems that the two sides didn't really match up and so it was very much favored exposing it in a band if you're doing two column work to arrange in your software to hop the gutter as it as it was called do it all in one go do a bit of left hand column and a piece of right hand column at the same timeperhaps it might be a good idea for those of you who really don't know an awful lot about this to just step back a bit and say well how did type setting and printing come about anyway and I'm sure many of you out there know that it goes back hundreds if not thousands of years in the Far East particularly it was possible to uh make marks on paper for Chinese Japanese Arabic characters 100 hundreds and hundreds of years ago those of in Us in Western Europe often fall into the Trap of thinking well really printing was invented in 1456 I think it was by Johan Guttenberg in Ms in Germany and of course he produced the famous Gutenberg Bibles but I think in all fairness you've got to say he didn't invent it he rediscovered it perhaps he refined the techniques and so on so to give you some idea I've got here a sample of wooden type and if you look down here it's actually for a font called caslon which some of you may even have the modern form of under windows or on your Macintosh or whatever that's just a display case but it it shows in many ways the way you used to have to set your type you had to have a wooden frame as big as the page size that you wanted if you had type of this huge size fine you could even handle it with your fingers and put it in in exactly the uh layout and format you wanted most of the time you didn't have the luxury of anything that big that is metal type it's metal type for a font called uni there it's effectively a if you like a sort of French equivalent of the Swiss helvetica type it's sand serif very clean very nice but just look at the size of it this is 10o univ there can you imagine that if you wanted to set one of your your papers or your letters with individual pieces of metal type you would be down to having a wooden frame armed with your Nimble fingers and tweezers maybe if you're going to set at 10 point you'll be picking these up putting them in lines having to accept the fact if you think about it you're going to put paper on top of these and in make an impression therefore these have to be mirror images in order to look right when you take the paper off and typically if you're on a A4 size page you're looking at something 25 times the area of this all to be done by hand and all to be made absolutely perfect no typos no nothing and how do you space the lines well up at the top of this box there's some flat looking pieces in there and those are pieces of what nowadays will be called leading if they're made in lead type which I don't know if this is but anyway the leading is the line spaces that go between your lines of type so you have to insert all those you have to make the lines look even the next stage is to gently roll ink all over it not too much ink you don't want blocks to appear on your paper and you would then put it in a press very much like an apple press or a cider press it's got a great big plate on it it's got one of those screw threads with a big handle at the top and the idea is uniformly and under high pressure to press the paper evenly onto the ink type and then you take the pressure off pull off the paper and if you're skilled enough you've got a beautifully ined version of what's there in your type from guttenberg's time for the next 400 years nothing fundamentally changed what changed in all of this in the late 19th century was that the setting of metal type instead of having to be done by hand became mechanized methods were devised for using these type samples almost as Masters and been able to replicate copies of them in a thing called a typ caster and once you got lots and lots and lots and lots of type enough to do a whole version of The Washington Post or the New York Times or whatever then what you needed was some way of mechanizing the placement of this replica type that machine was invented in Germany in about 1880 I think or thereabouts by a gentleman called Mar mgala he founded a very famous company called mgala the parent company but the name of the machine that did all this was the liner type and as its name implies it set lines of type which is exactly what the newspaper industry wanted a little bit later a rival came along worked under different principles which tended to Corner over the years rather more the if you like quality end of the book Market rather than newspapers that had to be bashed off every single day of the week that was known as the monotype machine I'm not saying there weren't others there were plenty of others because um type was a a good way to make a good living so long as you had the craftsmanship and the skills and so on but the big two were certainly monotype and linotype like the General Motors and Ford there were others but those were the big big names well Apple and Windows yeah apple and Microsoft Maybe so that technology persisted for about another 50 60 70 years uh of just setting metal type and putting in a printing press pressing the paper down and and so on but the next big revolution that came along from about the middle 1940s onwards was the idea of having your type not as metal but as if you like photographic images of the characters so if you can imagine it inside a big light tight box you would have a on a circular drum a strip of film that's got the whole shall we say of times Roman upright at the top and then below it times italic all the characters times bold times bold italic not all worked on exactly this uh principle but it gives you the idea and certainly the one they got at Bell labs this sort of second generation type Setter was very much like this so I'll stick to this description you've got all of your characters in the current font available as photographic images what you've now got to do is to transfer the photographic image of the capital letter s shall we say from the strip onto a piece of photosensitive paper at the back of the box and remember this is all light tight so what you have is an unbelievably complex system of mirrors and lenses capable of either shrinking or expanding the master letter s to be either 36 point or 6 point or whatever you're wanting so the sizing was done optically by magnifying and diminishing and when you were happy that you'd got the letter s to the right size this all takes place in milliseconds I might say with great noise in the background then you say right right I'm ready to print my letter s so you would fire a strobe a strobe light would take the image of the letter s focus it down through a fiber optic bundle and the fiber optic bundle would be positioned just above the photographic paper at the back at exactly the place where the letter s needed to be imaged and then you'd move the bundle a little bit off to the right and then maybe you do a lowercase letter A next door to it so if you can probably imagine it was mechanized but it wasn't Ultra fast you've got a lot of things to move a lot of synchronization to do in firing lights and so on but nevertheless it worked and monotype and linotype had Machines working on this principle but they weren't the only ones and the interesting thing is that at Bell labs in the early 1970s they actually had one of these true phototype Setters it was made by a small firm called graphic Systems Inc of Massachusetts so forgotten to ask Brian exactly how much it cost but is I would reckon tens of thousands of dollars armed with that the computer science section at Bell Labs developed an extended text formatting system called trof which was capable of type setting to this machine I should say that the GSI was just about computed drivable as it was supplied to you the idea was to drive it completely off paper tape which you prepared offline however it was possible to buy some extra electronics it was not a dedicated mini computer even in those days it's far too early it was specialist Electronics you could buy which would enable you to interface a parallel port type interface um off a pdp11 shall we say and to drive it directly from a pdp1 and that's exactly what bell Labs did so there they were with a text format which had been developed in the late 60s early 70s there were a lot of them around there's one at MIT I think called runoff Joo sna at Bells developed one called rough and what happened in those early days was if you ran ordinary rough normal rough as it were just to a text formatter that wasn't a types Setter again and it's back at this sort of quality that was called enroth okay but if in the future you wanted much better quality material you would have something developed called trough and TR standing for types Setter version of rough was capable of producing properly types set output looking just like this that's the way it was at Bell Labs from about 1972 to about 197 79 there was tough and its two pre-processors a guy there called Mike lesque developed one called tuble for typees setting tables you wrote in a tabular language how he wanted your tables um laying out and tubel's job was to translate your highlevel spec into lowlevel t- code and to squirt it into t- Ruffles Brian and his collaborator at the time Linda cherry wrote another one uh which was really just made your jaw drop open to actually typet mathematics and it's hard for me to believe actually even now that with the aid of eqn as of 1972 vintage you could very easily types set a page of that sort of complexity this is actually from a little bit later on it's University of Nottingham examination cover sheet but that the sort of mathematics type setting you could do with using eqn to drive this t- off program by the end of seven or eight years of very very heavy usage the poor old GSI cat was on its knees and uh Brian and his gang decided that they just had to have a replacement and by that stage the third generation had come along the third generation of types Setters actually used mini computers and actually represented the fonts as data inside a computer file the only problem then is well it's all very wonderful but how do you actually get an image how do you make an image how you know what do you do well in the third generation typically you used a very high Precision cathode ray tube to image your characters of course as you all know cathet tubes were the technology for televisions for many many years and yes you could make a catheter a tube you still can to be very high Precision indeed and by high Precision I'm talking about shall we say 972 dots to the inch now if you're going to make it that high Precision you are certainly not going to have a widescreen TV at that sort of resolution not in those days anyway this was where the actual Imaging took place and here there is a cartridge holding the silver bromide uh positive paper or your photographic negative film and underneath this black cover is the high resolution cathet tube somewhere yes is a picture of the highrose cathet which is like a letter box and essentially yeah and essentially what happened was you exposed the bromide paper in bands and and when you finished doing the current band you mechanically moov the bromide away from it if you wanted to do superscript and subscript work just by as it were dodging and weaving within the letter box that could be done without moving the bromide and that was very much favored mechanical alignment problems were a big big issue you had to have very high quality gears with no backlash because if you're setting a two column newspaper paper and if you decide to do it by doing the left hand column first and then winding back the bromide to the top and doing the right hand column next you didn't half uh come across alignment problems that the two sides didn't really match up and so it was very much favored exposing it in a band if you're doing two column work to arrange in your software to hop the gutter as it as it was called do it all in one go do a bit of left hand column and a piece of right hand column at the same time\n"