The dark intelligence of Scribus, the voodoo sides of ConTeXt, plus excitements and disappointments in html2print
(not finished yet)
P: I don't know if someone told you about this?
M: Yes, I especially know the text in the back.
P: Yes the text from Femke. Yes that's 2009 and at that moment it was the time before the LGRU and it was after several... but if we want we can pin that down more precisely... several attempts by OSP and Femke seperately, but mainly by OSP, to use Scribus for long texts, for books. And it was painful. I do not remember at what moment precisely... But Andreas VoksXXX, the Scribus dark intelligence, [laughter], has admitted at some point... I think it was around this time, before LGM 10... For me time is a bit divided between before and after a the LGM 10 in Brussels.
M: Was that during the LGRU, or before?
P: LGRU started at that moment, if I remember well.
P: So after the really beginning of OSP in 2006, 2007, 2008, many works in Scribus... I haven't found many examples [in the bookshelves in the Speculoos bookcases]...
And this is after I think, this is more like 2012. It's probably Scribus as well, made by Ludi. We were always in and out with Scribus, but it was really painful to work with long documents, because at time it was a very bizarre behavior of Scribus, growing exponential in size with the page count.
M: The more pages there are, the bigger the size of the file?
P: Yes, very bizarrely, because there is a kind of a loop inside that process. And so for that kind of stuff you are really obliged to divide the Scribus document. So with Femke we decided to go elsewhere, which was TeX. I don't remember how much text describes how much it has been super painful. It was even more.
M: There is a sense of painfulness, yes but I believe you it was more. [laughter]
P: But let's say, at least with Scribus it was more and more slow, and in the end it was a half gigabyte file to work with, but you are never stuck to the point that it's not giving output. You feel it's stupid and the slowness is painful, but you are not really stuck. You are with all breaks on, but not stuck.
And here it's compilation, so it's the same with html2print, and at some point your file is not compiling, and all the stuff you know... For me, and so for Femke, it was the first experience being so frustrating. To go to Hans HaakeXXX in the Netherlands, to ask... And clearly for me it was super traumatic as an experience. Because also at some point, my own limitiation in abstractions with the language, computer language I mean, was really blatent and in many occassions I have been obliged to let Femke alone. It was also very painful to see that. And you don't know what to do.
M: Because it was you and Femke working on it?
P: Yes yes. So the classic but frustrating for both sides, when the developer part is split. At some point Femke asked some help from Michael, but Michael was only helping on the principles of extreme computing and like that, but not in the core problems that we had. So it has been solved mainly by Femke alone.
M: Do you remember what the problems were? There were multiple problems right?
P: That voodoo stuff with TeX. What is bizarre with TeX is that it needs multiple loops to finalize a compilation. So it bizarrely processes a file and reprocesses it with stuff that it has gained in the first iteration. And some stuff goes voodoo. Some stuff are breaking and you see randomly there is stuff that go outside the boxes, that is not really a big problem, but also all the indexation system that has gone really another way. In the end it was beautiful but not the way we wanted to do it. And I do not remember the details, because it has been a very long time since I read the text from Femke...
P: [browsing the book]
P: And also it was really frustrating to see how we hang on the fact that it was not possible to change the font. Each time we changed the font it just exploded.
M: It didn't compile?
P: Yes. Or in a crazy way, or... So many dimensions of unpredictable. I think it has provoked my reluctance to go to other kinds of compilation based ways of working.
M: But you came close here, right?
M: [points to balsamine 2013-2014]
M: This was the first one that was made with HTML and CSS, but also many other things, I see Scribus, Python, Graphviz, Ghostscript.
P: Yes. It was made after this one.
[points to Balsa 2012-2013]
This was made in Calc.
M: It's super nice to see this one next to that one, first because there is already experimentation going on with different tools then Scribus, but also the small block elements come back in the way that the text is being placed in the page.
P: Yes yes.
M: What I wanted to ask... As far as I understand with the CSS Regions, they are used for two main things: one is to preview pages in the browser, how it's used in html2print; and second, to have predefined placement of text flows or images. So another flow, a second flow, can flow around it. Which is still something, this multiple flows on one page, that is still not possible to do these days.
P: With Paged.js.
M: Yes, but also not with just CTRL+P or other tools like Weasyprint. Either you need to really hack it and do it manually, but to let text flow around other objects is still not there. Which is ridiculous and interesting at the same time I think.
P: Yes, because it is considered too much "canvas", like PageMaker, Quorck Express XXX and Adobe InDesign. The way to link blocks freely, or pseudo-freely.
M: Because it breaks with the understanding of a page as a single flow.
P: It's the SGML approach. The first, eh, well for me the first document oriented design. Flows for documentation. Manuals and stuff like that. So the really old, super heavy duty environments to document technical stuff. And eh, what is the name of that super awful software? Because in the eighties, no early nineties, I was working in a production company and we were forced to win some competitions to use a super heavy system that has its own operation system to layout the stuff. And it was HGML XXX based, so really with one flow, but all references, cross-indexes and stuff like that. It's a name like Rosetta, but not Rosetta... Ventura!
M: Ventura, oke!
P: Because there was a huge market for that kind of stuff, there were also Xerox things and all these stuff. So it was a bit pre-HTML and pre-Word also. With the idea of text processing based on tags. And also TeX, TeX was before in fact. And also in TeX you only have one flow.
M: Because it comes from an academic background, no? Or from an academic context.
P: I think there are many papers written on that, on the fact that in academics you do not do multiple flows, because then it will break the consistency of it.
M: The single author?
P: Also the single dramaturgy of how you develop an idea and prove it, etcetera. A bit how judgement in law, you would not have multiple flows, you only have one and that is the law! [laughter]
M: You don't have multiple judges.
P: No no, no multiple versions. That's really interesting... That multiple flow stuff, comes probably from newspapers. Multiple authors, multiple stories, compilations of stuff.
M: Yes. And it's currently possible to make something like a newspaper layout in HTML and CSS, but it becomes difficult when you go to a next page. And to let it continue flowing.
M: In the Making Of text here it speaks about trying to step away from the tracks of Gutenberg. Or to take a small step away from the tracks of Gutenberg and a specific way of thinking of a page.
P: [laughs]
M: Do you remember that?
P: Yes.
M: It is interesting to see that stepping away is happening from Scribus in this case, which already is an act of stepping away from certain ways of doing. How would you describe a page made in the legacy of Gutenberg? What does this mean?
P: Wow, the Making Of text is really high level abstract, even in French. It is probably partly linked to the Gutenberg traces in the Verbindingen publication, that Femke and me made with ConTeXt. I'm not sure if Knuth, the person behind LaTeX, refers to Gutenberg directly. But if I remember well, Hans Haagen XXX, the person behind ConTeXt, was referring to that. And you feel it everywhere. So the understanding of Gutenberg's legacy is probably infused with that.
M: So the fixed margins that are always the same...
P: Yes, and were quite difficult to change in fact. We were fighting with that. Then the font... Many things were quite resisting any change in fact. Like it was hard coded.
M: Headers?
P: It's not Gutenberg in fact, it's an idea of how a document should look.
M: The biggest change after coming from Scribus in combination with other tools, to Calc...
P: Yes Calc, Scribus back with some frustration, then ConTeXt. And there is a bit of Scribus there too. It's always a bit there. And in Balsa 2013-2014, the real excitement was Graphviz and stuff. It was clearly that. It was for us more interesting and more rooted to the research, to the inside. But we also had to deal with the desires and frustrations of the clients, for legibility and stuff.
M: In a way Gutenberg is also an approach, no? It's clear that it's all about legibility. And it's clear it has certain traditions that you don't step away from, you do things in certain ways. But here [???] and here [???] the approach was very different. And I'm curious if we could articulate what the difference was in approach?
P: I think, for me when I have seen it taking shape, which was not too far from here actually, it was at Ludi's place. What was really exciting was to really code that. To see the flow and to saw, oke, break that in this and this and this. That kind of stuff was not possible to do in other stuff. And immediately there were also frustrations when it did not work, when it breaks things, but it was like a [wow]. There is probably inside the excitement of making things in another way.
P: What was super exciting in this stuff, was the fact that you feel the flow even if that flow is articulated or broken in a very specific way. And also to know that it was a flow. Or it's a flow that is guided through a quite detailed secrecy. At that moment it was feeling like, okay, we are there, yet it's only the beginning, and now some of our problems are behind us.
M: You mean some of the problems with Scribus and ConTeXt?
P: Yes. It was then possible to have at the same time a flow, with that practicality of efficiency and being able to lay out it in a different way for the web, etcetera. And at the same time it's something that is not looking like a flow, with all the clumsiness we want or need, and with the blocks, without the painful approach of working with a canvas to pick every part, like we did with Calc... Which is like layouting stuff in Photoshop... Working with HTML for print for Balsa 2013-2014 was super super exciting, I remember. And yes after, we could see that, for me, and I might be wrong... I never felt again that kind of excitement... All things afterwards were more disappointing, because the class with the standards, the fact that the browsers at that moment were so full of promises and not the kind of dissection we have seen afterwards with Google optimizing and stuff, removing possibilities that seem crazy.
M: Because Google removed CSS Regions from their browser, and thus removed the possibility to work with multiple flows and to position these flows on a page manually. When the single flow is the default, and your material is laid out automatically in a certain order, you forget how things are flowing into a page. You forget about the underlying fragmentation, the blocks, the flows, ...
P: Yes yes.