Some milestones in TeX history

The initial release of TeX was introduced by Donald E. Knuth in the early 1980s. The paper is called “The texbook”: Knuth, Donald Ervin, and Duane Bibby. The texbook. Vol. 3. Reading: Addison-Wesley, 1984.

In 1986 the LaTeX extension was programmed: Lamport, Leslie, and A. LaTEX. “Document Preparation System.” (1986).

The first description of the Lyx GUI was published in a Linux journal. In contrast to TeX and LaTeX before, Lyx was not focussed on university needs but had the enduser as a target group. Quill, Ulrich. “Introduction to LyX: Make working with LaTex easierusing the WYSIWYG editor LyX.” Linux Journal 1999.57es (1999): 6. https://www.linuxjournal.com/article/1355?page=0,2&%24Version=1&%24Path=%2F&quicktabs_1=1

A modern backend for the LaTeX language itself was introduced in 2005, called Xelatex. It is basicly a pdflatex program with improved font-encodings which works fine under modern operating systems. Kew, Jonathan. “About XETEX.” (2005).

The most obvious feature in the history of TeX is the long development time. From the initial TeX release until the current software around 40 years were gone. So TeX can be seen like UNIX as some kind of universal innovation which has evolved over the time. The first version of the UNIX operating system was modified heavily until his today iteration which is Red Hat Enterprise Linux, and the same is true for TeX. The first TeX version could be started only on commandline mainframe computers. A modern recent implementation has a GUI, runs well under any operating system and produces a modern PDF layout as default. Tex itself is not very mainstream compatible, but the underlying Bibtex format is well known. For example, the WIkipedia article about Bibtex has more traffic then the article about the endnote software. So perhaps some users who are not familiar with LaTeX are using Bibtex as well?

The most interesting aspect in TeX is, that it can be combined easily with modern technology. For example, the pandoc formatter which is a markdown parser uses for it’s pdf output an underlaying LaTeX backend.

What the future will bring is easy to explain. The latest development in academic publishing is called Open Access which is on a technical side a fulltext search engine like Google Scholar. With such a search engine it is possible to find a certain paper very fast. So the future of TeX will be in online publishing of PDF papers. The arxiv project for example, is based entirely on TeX. With Lyx it is possible to extend the effort into a mass market.

The main reason why TeX was a success has economical reasons. In theory it is possible to use any other wordprocessor for formatting an academic paper for example MS-Word or Adobe Indesign. But with TeX the amount of time which has to be invested by the author is lower. That means, if the bibtex database is there, the LaTeX template is ready and the images are available in JPEG format, the final pdf paper can be rendered very easily. It is way more faster, then drag and drop all the information into a MS-Word document and adjusting manually the floating images. Sure, the TeX ecosystem is a bit complicated to grasp. But it has improved over the years and todays version of Lyx has everything out of the box what beginners in typesetting will need. The most interesting feature in Lyx is not the sophisticated layout but it is the feature to export a text into a plain-ascii format which includes the bibliographic references as well. That means, even somebody doesn’t like LaTeX he can write with Lyx nice looking pure ascii files. Left on the screen he has the outline editor, and after he has typed in the text he simply exports the file to plaintext ascii. This bypass the xelatex backend and it works quite nice.

Advertisements

Early history of TeX

The beginning of TeX can be dated back to the late 1970. Google Scholar finds for the author “Donald E. Knuth” from that area some papers, which are introducing TeX and metafont. The interesting aspect about the first papers of 1979 and later improvement is the very low speed the development itself. That means, the first version of TeX was presented in 1979. It takes until 1984 until Bibtex was ready, and in 1989 some type3 fonts were released. That means, a minor improvement took not weeks but years. And this slow development speed is an ongoing process. For example, it takes many years until the first version of Lyx was programmed. Literally, the TeX community is working since mid 1970 on his system and every 5 years or so, they have invented something new.

The first question which has to be answered is, does TeX makes sense? I think the answer is yes. The productivity is higher then with MS-Word and it has a high output quality. The next question is, how can we improve the current software. I would guess, future improvement have to build on top of Lyx. Lyx is the most important frontend right now, it uses an underlying bibtex and latex backend for generating the output. I would guess the most room for improvement is the sourcecode itself. Rewriting Lyx and LaTeX with modern C++ would be a nice task for future students.

The year which can be called the birthday of TeX was 1984. That was the year, in which the TeXbook was published (written by Donald E. Knuth) and in the same year, the LaTeX manual and the Bibtex draft was presented. The PDF format was unknown in that area, also Type1 fonts and the Lyx GUI. These developments came later. The basic idea behind TeX was to replace mechanical typesetting machines. The idea was to generate a printable document with a computer. TeX can be seen as direct competitor to Phototypesetting.

A long journey

According to the timeline it is a fact, that between each iteration of TeX many years were in between them. TeX was programmed in 1979, LaTeX in 1986, pdflatex in the early 1990s, Lyx in the late 1990s, Google Scholar in 2005 and so on. Each step adds an important functionality to the work done before, but the overall workflow took around 30 years. It is very likely, then in future improvements will come also in a period of 10 years or more. Why?

One possible explanation is, that the computersoftware TeX is used by people, and if new sort of users are become aware of it, they are programming an extension. The long years between each period has nothing to do with the software itself, but it’s adoption from a certain usergroup. Lyx is perhaps a good example. The Lyx community is different from the original Tex and LaTeX community. They are not familiar with the commandline or Emacs, but with MS-Word. So they have programmed their own TeX which looks like Word, but is more powerful.

Perhaps the future Facebook generation is not familiar with desktop applications, and they have a need for a Lyx version which runs in a browser window. I don’t know, because it is unclear what the future will bring.

Productivity in printing history

The first gutenberg printing press in 1500, the Koenig&Bauer fast-press of 1811 and the Linotype machine from 1930 have something in common. They were all used for printing books and newspapers. They can be classified according to their productivity, that means how many man-hours it takes until a result can be seen.

The printing process was not done by a single person. It was not the case, that Gutenberg alone has used the machine, and also a steam driven printing press of 1850 were always put to work by at least 20 workers. They can work on different steps of the workflow at the same time. This is called parallel executation like in an assembly line. It helps, to get the result faster done. But, it didn’t changed the productivity. If the amount of work which is needed to get a book printed around 1000 hours, then it is possible to use 10 workers with each 100 hours, but if we want to pay the salary for the workers it is the same amount like letting one worker work for 1000 hours.

The main difference in the technology was an increased productivity. The number of invested hours reduced over the years. This results into a cheaper product. The costs for making a book are smaller on a Linotype machine compared to the gutenberg printing press of 1500. But the disadvantage is, that modern machines are more difficult to built. The first gutenberg press was build of wood, while a modern Linotype machine is a complex mechanical device.

So called printing museum are available worldwide. What is missing are the negative aspects of printing. Usually the printing press is described as a beautiful machine which has made the life easier. But in reality, printing was criticized from the beginning. The ideology in doing so is called Luddism. Luddist asks what a printing machine is replacing and what society will lost from reading too much.

Is LaTeX outdated?

Every technology is under pressure, and it makes sense to even ask if the LaTeX software (invented by Donald E. Knuth) is perhaps outdated. But before doing so, let us get the facts right. Around the 1930s the most modern form in typesetting was a Linotype machine. That is a very fast printing press which reduces the effort in making a newspaper. In the 1960s Donald E. Knuth developed for a mainframe computer the first version of TeX, which was extended in the 1980s to LaTeX. In the late 1990s the Lyx system was programmed which extends LaTeX with a GUI. In the mid 2000s the new backend version of LaTeX called Xelatex was released.

And now we can answer the original question. The original TeX software from the 1960s is today outdated. It is no longer useful. But the predecessor of TeX has the same name, it is some kind of better Tex. The latest iteration contains the Lyx frontend and the Xelatex backend and is able to generate pdf documents on the fly. Calling this technology outdated isn’t possible, it is the most advanced form of typesetting ever invented. In contrast to old school Linotype machines it works entirely on a computer and in contrast to MS-Word it is Open Source and has a semantic markup language. Perhaps, one day, the combination of Lyx and Xelatex will be outdated too (like any other technology before) but today it is state of the art. That means, it is the perfect choice for producing scientific documents which has lots of figures and citations.

Peer review with printed documents

Often, the peer-review process is described from a content level. But let us calculate the simple paper costs for doing such a quality control system. Suppose, the class consists of 20 students, each of them writes an essay with 10 pages. Peer-review means, that somebody else is reading and commenting the paper. Each students prints out a copy of his paper for his classmates. The total number is: 20 students x 10 pages x 20 copies = 4000 pages. No it is not a joke, a peer-review in a small class needs such amount of papers. Printing out one page costs around 0.15 US$. The overall costs become 600 US$. Who want’s to pay the price? Right, nobody. And that is the reason, why no peer-review is take place. Peer-review is a theoretical concept not used in reality, because the price of printing out a copy of a paper is to high for most students. And that are only the simple printing costs. None of the papers was read, this has to be happen too.

Now imagine that we want to scale up the peer-review process. We are repeating the peer-review task ones a month, and for 1000 schools at the same time. What will happen is an explosion in the costs, it will result into million of US$ how have to be spend only for distributing self-written essays. It is not possible to make the process more economically, because the costs of 0.15 US$ per page are fixed. It is the lowest possible price in the copyshop. The only thing what is possible to reduce the cost is not doing any peer-review, and this is what took place in the united states.

History of the printing press

The video shows the workflow of the printing press which was common around 1900. At first, the letters were set manually into the plate, and then the printing process starts. How can the people in that time mastering the complicated process? In the demonstration, the entire workflow is demonstrated by a single person, that is historically not correct. In reality, the workflow was divided over many workers at the same time, like on a assembly line. This results into a speed up.

For example, in the youtube video it takes around 10 minutes until the printing was possible. It is a linear process from the beginning to the end. Most of the task can be done in parallel, in a similar way like a modern cpu can execute parallel task. Worker 1 is preparing page 1 with mechanical letters on the board, at the same time worker 2 is preparing page 2 and so on. If around 100 workers are available the preparation until the machine is ready for the first printing is very short. But, the disadvantage is, that more workers cost more money. And if 100 workers are working each 1 hour, it is the same price, as if 1 workers runs for 100 hours.

That it means, even it was possible to parallelize the work a simple “hello world” printing costs a huge amount of money, which prevents, that a book can circulate freely. The first printing press was not an enabler for knowledge distribution, it was more the opposite.

Limits of the steam powered printing press

The first steam powered printing press was invented in 1811 by Koenig&Bauer. It was later improved to the Rotary printing press in 1844. It can be called a revolutionary device, because at first time in history it was possible to print large amount of books and newspapers. The main problem with the printing press is, that it is a dead end. It can’t be improved anymore. Todays printing press are often computer controlled and have an increased performance but their main problems are remaining the same. It is mainly a mechanical device which is very costly. That means an average printing press is made of steel and needs sheet of paper. The costs of the device are fixed, that means it is not possible to print a book for under 1 US$. Even on the most modern printing press the costs are much higher.

A second problem is, that the printed book / newspaper is located physically. That means, the new issue of the new york times which was printed in New York, can’t be read in Europe. Because the sheet is produced in Newyork and that is far away from Europe. That means, a book can only be read around 50 km around the machine and has to be transported with an airplane of a ship over longer distances.

A printing press was only the first step in media revolution. It was a starting point for other inventions like the telegraph and microfiche devices. It is surprising, that even today, the printing press is called a defacto standard in publishing. It is not, because the system works so good, it is because of a collective anti-technology thinking. That means, from a moral point of view, a steam powered printing press is fair, because it prevents an information explosion. The bug of the system, that the new york times can’t be read in europe is explained as a feature, because the people are protected from reading to much.

I think we must evaluate the invention in his time. For the year 1811 a steam powered printing press was advanced technology. That means, it solved a lot of problems, which were open in that times. So it was the right invention to the right time. But it makes no sense to imagine, that the year 2018 is equal to 1811. Between them are more then 200 years. Today, a steam powered printing press is obsolete. It is not an invention it is outdated.

The reason why I’m talking about the printing press is because from a certain point of view, it has neverthe less something innovative. If somebody is putting the steam powered printing press in a museum and describes it as the past, then it become the future. Because, if the printing press is something of the past, how does the machine look like, who has replaced the steel monster of Koening&Bauer?

To make the point clear, people have used steam powered printing press for over 200 years. They have used these machines for every kind of books and newspapers. They have done so, because they were not aware of a better alternative.