Some comments on ‘The Digital Orientalist’ blog

It is refreshing to see a scholar of ‘Oriental’ studies concern himself with the representation of texts, and visual aspects of the written word. By and large, looking at texts has held little or no priority in most of the humanities (a startling fact that would merit discussion), and so I was interested to see Cornelis van Lit, writing as ‘The Digital Orientalist’, announce a series on Digital Printing of Arabic.


About: Digital Printing of Arabic: explaining the problem

In the first instalment, van Lit proposed to explain fundamental problems of the topic. Unfortunately, some of the exceedingly short ‘explanations’ got bogged down themselves in the complexity of the matter, and others introduced further and unnecessary confusion. Here I am trying to address some of these issues and correct some errors.

Van Lit sets the tone of his piece with an opening paragraph that pits the Arabic script against the ‘digital world’, underlining that the Arabic script was developed a long time ago and that it was primarily written with pen and ink on various supports. Whilst this is true, it doesn’t add anything, for the same can be said about the Latin script, and the Chinese script, and pretty much every other script humankind ever invented. And most are as old, or trace their roots even further back. Suffice it to say that such mystification of Arabic is both wrong and counterproductive.

In the paragraph on printing, lithography is mentioned prior to foundry type, even though it was invented some 400 years later (van Lit uses the term ‘movable type’ for foundry type. I consider this somewhat imprecise as other typesetting technologies may also be called movable type, even if they do not use metal sorts). Sentences such as ‘printing became much more popular using the technology called movable type [emphasis in source]’ lack clarity and factual content: where, and when did they become more popular? Surely not in the Arabic script world, where lithography was an instant and huge success in the second half of the nineteenth century, after foundry type had not been embraced for the preceding 400 years.

Turning to foundry type, we are shown a single image of an abysmally bad fount, printed in the late nineteenth century in the Netherlands. This is misleading, as at the same time printers in Beirut, Cairo, and Istanbul had developed much better foundry type that had overcome such alignment and consistency problems. The central problem that the example highlights, and this seems to be lost on van Lit, is that the problems are not inherent in the technology, but demonstrate that western type founders lacked an appreciation of Arabic letterforms. They failed to reproduce the proportions, stylistic characteristics and general aesthetics of Arabic script not because the metal could not be shaped to the will of the punch cutter, but because they had not learned to look at Arabic with the same acuity as they did look at Latin – arguably a shortcoming that is shared by many a learned scholar. Whilst van Lit’s observation that the technology was not particularly well suited to the properties of Arabic is correct, it fails in explaining correctly where and how the technology imposed its constraints – the central theme that the article is aimed at. Here, I was also surprised to read the term ‘letter block’ in the attempt to explain some of the morphographical differences between the Arabic and the Latin script, for as far as I know, Thomas Milo and I coined this term in the discussion of DecoType’s analysis of Arabic script grammar. To find it here, out of context, and without attribution, could either be a sign for the decisive success of our terminological contribution, or rather poor attribution.

Van Lit then jumps to the representation of Arabic on computers, skipping roughly 100 years and numerous technologies that all had tangible, and often profound effects on the appearance of the script. The summary explanation that most typographic technologies originated in the west, and were thus shaped around the properties of Latin is not wrong, but not very original either. In its brevity it does not help to understand what exactly happened in the numerous stages of this evolution. This is important, for van Lit’s following conclusions and attempted explanations all fall short of identifying the actual problems. We cannot talk of ‘computers’ and ‘the digital environment’ in such general and blurry terms if we want to understand how technology influenced the representation of texts. If an incompetent ‘tattoo artist’ (or probably rather the client) uses the default system font in an application that does not support right-to-left composition (or has it turned off), the issue is not with the technology, but simply his ignorance of Arabic. Of course the offence is exacerbated if this happens at an airport, or on the poster of a cultural institution, but it’s ill-informed to label errors of use as flaws of technology, for this distracts from the actual shortcomings technologies had, and still have.

Turning towards encodings, we come a little closer to the gist of the matter. Varying encodings have indeed been a problem of early digital text storage, as different vendors used different codes to denote the same letters. This meant that documents could not be exchanged across various applications and platforms, as the same code could mean one letter on a PC, and a different letter on the Mac (this is simplified, but sufficiently detailed for the present purpose). Van Lit’s demonstration of searchability problems are indeed related to the encoding, but not exclusively so: it is no coincidence that the illustrations are all based on PDF documents, for searchability problems of Arabic are specific to the PDF format (and bad fonts!). Similar illustrations could not be made with a standard word processor as have been in use since the early 2000s, and it is very probable that different PDF readers would yield different results. So yes, encodings are related to the representation of digital Arabic text, but it is far off the mark to say that they are responsible for typographic issues of Arabic in the digital environment. Indeed, without encodings, there would be no digital Arabic at all.

Unicode, the next topic van Lit raises, is the current standard encoding, and has been since Windows 2000 (with the Mac platforms trailing by a few years). That van Lit begins his discussion of Unicode noting that ‘solutions to such problems have been proposed, but not widely or accurately implemented’ is severely misleading, as today it has become very difficult to write any kind of digital text that is not encoded in Unicode. Indeed, Unicode was developed because of the aforementioned encoding issues, and has made enormous progress in overcoming them. Not only does Unicode-compliance of fonts and software solve the encoding problems van Lit illustrates, the Internet would be inconceivable without Unicode, and all of our day-to-day exchanges, Arabic, Latin, or Emoji, are based on Unicode. Whilst Unicode is far from perfect, as van Lit correctly illustrates, it can hardly serve as a scapegoat in a discussion of international computing.

And so, an uninitiated reader of this piece gets away with a diffuse sensation that somehow technology, and print, and the west, are to blame for some problems of Arabic computing, that he still does not understand. How this could help to make amends to these problems remains open to the imagination. Van Lit’s conclusion that ‘print’, ‘movable type’ and ‘especially Unicode’ introduced these problems, betrays half-baked ideas and a poor understanding of the topic. His closing remark that ‘not only can we present [Arabic] on a screen and print it on paper in only a rudimentary fashion, we often encode it such that it is not flexibly searchable and therefore has limited repurposing possibilities’ could not be more wrong.

About: The Mushaf Muscat

In a further instalment of the series, van Lit discusses a recently launched online edition of the Qur’an: the Mushaf Muscat, jointly developed by the Ministry of Religious Affairs of Oman, and the Dutch company DecoType. Again, a very impressionistic portrayal of the subject is offered, with glaring inaccuracies and a lack of research. In addition to the issues pointed out in comments on the blog, here I want to raise just a few points that relate to the technical aspects discussed above.

In his comments on the design of the Qur’an, Van Lit writes that ‘the calligraphy does not pretend to represent a particular historical reality. This is a new, modern rendering of the Koran, which follows a self-developed naskh for the body of the text and a kufic script for the headers.’ How van Lit has arrived at this conclusion remains opaque, and it does not hold true. DecoType’s Naskh typeface is a meticulously researched model of the Ottoman school of Naskh penmanship, precisely delineated by strict selection criteria. Indeed, I would be hard pressed to name any Arabic typeface as pure and deliberate in its choice of models as this, a true thoroughbred if ever there was one. The insinuations that the adjectives ‘new’, ‘modern’, and ‘self-developed’ suggest remain inaccessible to me, which may be just as well. And whilst I know less about DecoType’s Abbasid type (Kufic does not really describe much, especially not so in type), I would be surprised if it was based on anything less rigorous than the company’s Naskh.

Van Lit’s assumptions about DecoType’s technology are unfortunately really just that: assumptions. When writing that ‘DecoType is able to circumvent all of those issues introduced by movable type and unicode [sic]’ in an approach that he considers ‘fairly simple’, van Lit not only gets it wrong, but also denigrates some 30 years of work at the forefront of Arabic type technology. Simple? Circumventing Unicode? Far from it: DecoType’s technology offers a fully-fledged rendering engine that is Unicode-compliant and tailored towards Arabic script characteristics at the same time. Yes, these are fonts, and Tasmeem, the software that is the current commercial embodiment of DecoType’s technology is available for Adobe’s InDesign – nothing short of the global market leader for print publications.

In conclusion, the best part of these articles is that they are written from the perspective of an outsider with an interest in the subject. Whilst such commentary used to be common in a field as narrow and poorly defined as typography and print, it is becoming less acceptable as the discourse matures, and something actually worth calling ‘a field’ emerges. So should we stay tuned for more? Yes, sure, for talking across discipline boundaries is sorely necessary. But please make sure it will be better researched next time.