| Wishful thinking or thinking ahead? Envisioning the next generation of SGML editors | Table of contents | Indexes | The Evolution of Sun's AnswerBooks | |||
| Guittet Christian |
SGML, still a cutting-edge technology? |
| Christian Guittet |
Abstract: |
| Was WasSGML born too early? 10 years ago, very few people understood its potential. But, now-a-days, surfing on the Internet wave, a number of technologies have become trendy. Buzzwords like virtual reality (VRML) or active content ( Java, activeX) have become fashionable. How are they related to Java, activeX) have become fashionable. How are they related toSGML ? Is it not the ? Is it not theSGML approach that paved the way to such advanced technologies? |
| This paper explores these issues, examining the results of a number of research projects launched by Eurostat to review cutting-edge and emerging information technologies and to evaluate, from a user's point of view and on the basis of concrete applications, how they could help Eurostat fulfil its mission, which is to provide the European Union with a high quality statistical information service. |
SGML , still a cutting-edge technology? |
| Documents have been published for thousands of years. Why can't we just carry on? Why should we change our methods? Why do we need Documents have been published for thousands of years. Why can't we just carry on? Why should we change our methods? Why do we needSGML ? Because of multimedia, you said? |
| Multimedia documents have been published for thousands of years: at that time, the buzz words wereilluminations andmanuscripts ; but they did also refer to the achievement of asymbiosis between several means of expression, i.e. several representations of real or imaginary worlds ? Why can't we just carry on? Why should we change our methods? Why do we need ? Why can't we just carry on? Why should we change our methods? Why do we needSGML ? Because of printing, you said? |
| Documents have been printed for centuries. Why can't we just carry on? Why should we change our methods? Why do we need Documents have been printed for centuries. Why can't we just carry on? Why should we change our methods? Why do we needSGML ? We don't need it, you said? |
| Or, to put it in other words: was Or, to put it in other words: wasSGML born too early? born too early?SGML , still a cutting-edge technology? This is the title of my presentation, but I should rather have asked: This is the title of my presentation, but I should rather have asked:SGML , at last a cutting-edge technology? |
| When WhenSGML was made an international standard, some 10 years ago, very few people understood its potential. was made an international standard, some 10 years ago, very few people understood its potential.SGML : Sounds Good, Maybe Later - surely that was more than just a joke, some people did mean it. But, now-a-days, surfing on the Internet wave, a number of technologies have become trendy: buzzwords like active content (java, activeX) or virtual reality (VRML) have become fashionable. How are they related to (VRML) have become fashionable. How are they related toSGML ? Is it not the ? Is it not theSGML approach that paved the way to such advanced technologies? |
| Indeed, there has always been a profound misunderstanding around Indeed, there has always been a profound misunderstanding aroundSGML . It was (and, possibly, still is) considered by many to be something very complicated and almost useless, some kind of a nuisance imposing on users to put everywhere markers that make the whole text completely illegible. But . It was (and, possibly, still is) considered by many to be something very complicated and almost useless, some kind of a nuisance imposing on users to put everywhere markers that make the whole text completely illegible. ButSGML does not deserve this bad reputation. |
| Let us start with an easy example, which we will look at in details: theECU yields . Since September 1992, Eurostat publishes data on the level, structure and volatility of ECU yields. To make a long story short, the yield is a function of the maturity, involving 5 different coefficients provided daily by theISMA ((International Securities Market Association)) , reflecting the bid-offer prices of a comprehensive range of financial institutions.: |
| Y = C0 + C1 .M + C2 .M2 + C3 .M3 + C6 .max (0, M-5)3 [1 =M =10] |
| If you know the coefficients, you have all the information - but not in a practical form. In this sense, the file in figure 1 contains everything you need (or need not) know on the recent trends of ECU yield. Please also note that the data format itself is obviously If you know the coefficients, you have all the information - but not in a practical form. In this sense, the file in figure 1 contains everything you need (or need not) know on the recent trends of ECU yield. Please also note that the data format itself is obviouslySGML -conforming: all of the mark-up has just been made implicit, using the full power of -conforming: all of the mark-up has just been made implicit, using the full power ofSGML , including extended features such as DATATAG. |
|
| URL:http://euopa.eu.int/en/comm/eurostat/research/subcom95.21/java/ecu/ecudata.txt |
| Yes, this file is an Yes, this file is anSGML file, although it does not look like it: no mark-up is to be seen, because it is hidden in the data itself. |
| I agree, it is difficult to read in this form; some calculation and formatting will make the information more readable, more understandable. A table showing a number of key yields, such as the one in figure 2, will indeed look better: |
 |
| URL:http://europa.eu.int/en/comm/eurostat/ecuyield/wr1.htm |
| Yet an even better 'layout' for this Yet an even better 'layout' for thisSGML file is possible. Static tables are no longer fashionable to-day: cutting-edge documents do have anactive content - an applet can format data "on-the-fly" and introduce interactivity. |
 |
| URL:http://europa.eu.int/en/comm/eurostat/research/supcom95.21/wr1-java.htm |
| By the way, just a reminder! Width, height, ... I fought against such attributes some 12 years ago when By the way, just a reminder! Width, height, ... I fought against such attributes some 12 years ago whenSGML was being defined, because they do describe the physical layout of a document, not its logical structure, defeating the very purpose of was being defined, because they do describe the physical layout of a document, not its logical structure, defeating the very purpose ofSGML . I have not changed my mind and do not like them any more to-day; they are just a weakness of HTML, which has unfortunately been imposed by the way browsers work. |
 |
| URL:http://europa.eu.int/en/comm/eurostat/research/supcom95.21/wr1-java.htm |
| But look at the layout of figure 4: the result does not look any longer like a traditional publication. Not only has raw data been turned into a graph while the But look at the layout of figure 4: the result does not look any longer like a traditional publication. Not only has raw data been turned into a graph while theSGML file shown in figure 1 was read, but the user can also interact with the document; he just needs to enter any maturity to be shown the corresponding yield. |
| This example demonstrates a new, dynamic approach to classical graphical representations of statistical data: marking-up a data file is a pre-requisite to bringing interactivity. |
| Even more innovative ways to visualise statistical information are possible and have also been examined at Eurostat. For instance, a large set of data can be synthetically represented as a colour table: applets can read marked-up files, associate colours to values contained in cells and dynamically draw the corresponding 'painting'; the user can then interact with the result to apply image processing techniques and find out the real information that was hidden amidst useless details. |
| For example, the core of the 'document' shown in figure 5 is a java applet that reads statistical data from an For example, the core of the 'document' shown in figure 5 is a java applet that reads statistical data from anSGML file selected by the user, processes it and presents the result as a 'colour table' which can then be processed interactively as an image. |
 |
| URL:http://europa.eu.int/en/comm/eurostat/research/supcom95.27/release/bin/classes/lot27.htm |
| Is this approach intrinsically limited to the statistical field and to the graphical representation of numerical data, you may ask? Certainly not! For instance, a marked-up text file can similarly been turned into a piece of edutainment. |
| Figure 6 indeed illustrates how a list of questions and answers, in different languages, can be 'formatted' as a quiz. To minimise file size and download times, all the mark-up has been hidden in the data itself, as permitted by the Figure 6 indeed illustrates how a list of questions and answers, in different languages, can be 'formatted' as a quiz. To minimise file size and download times, all the mark-up has been hidden in the data itself, as permitted by theSGML standard (the trick lies in using the DATATAG feature - it permits data to be used as markers, i.e. certain characters serve as data and markers simultaneously, depending on the context): the first element of the file is a question in English; the 'carriage return-line feed' at the end of the line represents the 'open marker' of the same question in French; the 'carriage return-line feed' at the end of this new line represents the 'open marker' of the same question in German; the 'carriage return-line feed' at the end of the line now represents the 'open marker' of the answer; and so on ... |
 |
| URL:http://europa.eu.int/en/comm/eurostat/research/supcom95.21/quiz.htm |
| Is that all? will you ask. The answer is: NO! Why stick to publications in 2 dimensions? VRML is popular now-a-days, isn't it? OK, then, let's use VRML is popular now-a-days, isn't it? OK, then, let's useSGML to make 3D interactive publications ... |
| A prototype has then been created, to create awareness around this new possibility, to assess user acceptance and to identify potential difficulties. A java applet, made available on the Web, allows a user to input a statistical table of his choice and to format it on-the-fly as a VRML file - which, by the way, could probably be described as a minimised VRML file - which, by the way, could probably be described as a minimisedSGML file. This VRML file is then visualised as a 3D bar chart, which the user can manipulate: zoom in, zoom out, rotate, slide, ... More generally, anyone can now 'see' the statistical information from different view points, select the most appropriate one and ... capture this specific view of the chart to publish it in a traditional, printed publication. |
 |
| URL:http://europa.eu.int/en/comm/eurostat/research/supcom95.21/barchartgen.htm |
| Is that all? Yes and no! Yes, if you consider only proven technology. And no, as some more promising technologies will soon become widely available: for instance, collaborative VR will shortly provide multiple users with a simultaneous access to a common virtual object. While the generalisation of some applications, relying on the availability of VR theatres, tabletop VR systems or liquid crystal shutter glasses, are still a long way ahead, Internet-based, immersive, interactive access to VR objects, using off-the-shelf products, can already facilitate both access to statistical information and communication between partners. |
| And there again the future will show that And there again the future will show thatSGML does have a key role to play... |
| Wishful thinking or thinking ahead? Envisioning the next generation of SGML editors | Table of contents | Indexes | The Evolution of Sun's AnswerBooks | |||