It is not often I read something criticising the way Microsoft works and find myself agreeing with all the arguments. However, Alex Schlup of DSK Networks wrote to a trade paper recently and explained how Microsoft maintains its monopoly, (well I don't agree with that I suppose), by controlling standards. The argument is not new but his explanation certainly highlighted the problem more than anything else I have read. This explanation, together with my recent experiences with Word 97, prompted me to ponder again the subject of standards in the present age of computing.

The use of standards has always been one of those recurring arguments. Whether to abide by standards, or break the rules, each take their turn in precedence.

The argument for standards is that it reduces costs in a number of ways. It reduces support costs because you only have to learn one set of rules. You can buy the cheapest software that supports your standard, or the most functional to your need. It is much easier to communicate, be it between software packages or from one place to another, if standards are adhered to. For example, it is because we use HTML that we can all browse the web very simply and cheaply.

The argument against standards usually boils down to the one that it stops innovation. When standards of software development are laid down programmers usually squeal that it takes away their individuality and that work is boring. That is probably true, but the upside for users is they usually get software that works developed faster! When standards are in the hands of committees it does take a long time to get change incorporated and this is a severe problem when the change is generally felt to be 'good'.

To give a real world example of the benefit of standards and what happens when we start to deviate from them I would like to draw your attention to bicycle racks! I live in Stevenage, England. Our cycle ways have been famous for a number of years, (c.f. the 60s’ film ‘Here We Go Round the Mulberry Bush’), and I do quite a bit of cycling. In places where there are a number of cycles stored, such as my current workplace and outside the railway station, you will find cycle racks that are built round a design that was implemented many years ago. Basically they consist of a parallel set of runners to place the wheels of the cycle in. They alternate between having a bike angle of 0 degrees to the horizontal and 30 degrees to the horizontal. The idea is that the handlebars of adjacent cycles do not interfere. This design is great but it depends on two crucial standards that bicycles were designed to meet. First the runners will only accommodate the wheels of a bike if their width was less than 1½ inches. Second, the width at the handlebar end had to be no greater than twice the distance between runners, otherwise they would interfere with other bikes. Now these standards lasted throughout my childhood and into my thirties. There were different types of bike - we had road bikes, sit up and beg handlebars, drop handlebars but essentially these two standards were met. Now look what has happened during the 4th and 5th decade of my life. In this age of designer products the age of the mountain bike has come. I find it very difficult to find a 'normal' road bike now! My last one came from one of those adverts in the weekends paper, which said they had a warehouse full of unsold bikes! Very good it is too. However, most people ride around on a mountain bike, despite the fact you are unlikely to find a mountain for a hundred and fifty miles from here. What does this mean. It means they have chunky tyres, which are too wide for the runners of cycle racks. The result is that people park them between the runners where at minimum it means you cannot park a normal bike in the adjacent runners and quite often they fall over taking up more space than they ever would. In the last few years I have noticed a trend for bikes to have very straight wide handlebars. God knows why but the effect of this is that you cannot place another bike within 3 runners either side of it without the handlebars fouling. When I try to get my bike in closer than this I have often come back to find it sprawled all over because owner of wide handlebarred bike cannot be bothered to extract it carefully, so knocks my bike over and fails to replace it. The overall effect of this deviation from standards is that these racks can only hold less than half the quantity of bikes they could in days gone by. At a time of increasing cycle use this seems pretty stupid to me.

Now that I have that off my chest lets get back to the world of computing! When we started computing at home we already had a standard for text on a computer. It was the American Standard Code for Information Interchange, or ASCII. OK, so Commodore altered it a bit with the version that came to be called PETASCII but it did not take much BASIC programming to convert PETASCII to ASCII. The ASCII standard has served us very well. If you wanted to export data from one program and import into another then using ASCII was pretty much the way of doing things. It is interesting to note that in the big iron days, when IBM ruled the roost rather than Microsoft, they used their own proprietary 'standard' called EBCDIC, pronounced ebsidic, which stands for Extended Binary Coded Decimal Interchange Code.

As time went on, and everyone started to use word processors and coloured screens and printers, it became apparent that there was a problem with ASCII. It could represent text but it could not represent formatting information. There was no way to say in ASCII that the text was bold, or italic, or was coloured red, or indented one inch either side from the margins. Various solutions to this puzzle were found. Sometimes codes that were not allocated to the alphanumeric characters were used. Sometimes, as in the days of Superscript, tags were used so that, for example, /lm10 meant set a left margin of 10. Although there were different ways to cope with the problem no standard solution appeared. Rich Text Format, (RTF), was an attempt but it never really caught on. The big three in word processors, Microsoft Word, WordPerfect and AmiPro, (now part of Lotus SmartSuite), all used different techniques and transfer of data between them meant special routines had to be written to translate the codes. Transfer to other types of programs, such as spreadsheets, was even more unlikely. We no longer had a standard as such.

As the years passed this became a real problem for business users, who have a greater need to move data between programs. Their solution was to standardise on a set of software from one manufacturer, who would support the transfer of data between their own applications. The Office suite was born and it did not take long for Microsoft to become the de facto standard. The other alternatives, WordPerfect Office or Smart Suite, were a bundling of products from, initially, different developers. When they were all under one roof the developers took ages to develop the essential interoperability, by which time Microsoft had won the war. This is why Microsoft became the number one Office suite. Not by devious means, as the DOJ would like us to believe, but because they recognised what was needed and provided it.

Microsoft Office went from strength to strength and by the time it reached version 4.3, which included Word 6, Excel 5 and Access 2, it was pretty damn good. I would say there were only two faults with this product. The macro languages behind each application were not common and an application itself contained everything that was likely to be needed. Excel did load some specialised tools only when requested but Word was just one big huge program. Some say bloated but in reality what is inside Word is needed. It is just that different people need different things so trying to take away something because it is not needed by one person is bound to upset someone else. The solution is software components which are loaded when required but, despite Microsoft promising this approach for a long time, this has never happened.

Getting back to word processing Microsoft had, by default, created the new standard for text transfer. What Alex Schlup objected to was that the 'standard' was in the hands of a developer rather than an independent body. The reason why I think he may be right is to consider what has happened since Word 6. Microsoft has subsequently developed Word 95 (internally Word 7), Word 97 (internally Word8) and Word 2000. Each of these versions has added more functionality, which some, including me, would say is dubious. Word 95 was OK because it stuck with the same text format as Word 6. However, Word 97 and beyond has created new formats with the inevitable problems of transfer of data between users of the differing versions.

Let me discuss Word 97, since I now have a fair amount of experience with it. I have moaned many times about its lack of capability of dealing with HTML properly, despite this being one of the lures for upgrading. I would now like to address another 'feature', that I have had cause to curse many times in my workplace. Microsoft have been trying to encourage group working where many people can work on the same document and each can know who contributed what. There is a feature in Word 97 called 'Track Changes', which was called 'Revision Marks' in Word 95. With this turned on you can track how a document is changed by various people. It can also be used to mark in the border where changes have been made. This is the feature I use because when I create an issue 2 of a document I want to draw the reader to the changes made. I do not want them to waste their time reading stuff they have read before. The way this feature is implemented is to store both old and new versions of any changes in the document. Now suppose you make a few mistakes and correct them in rewriting the document. Unless you take deliberate action you are going to have all the mistaken typing hidden in the document saved. You end up with a document that is full of junk. Do this with a few pictures and the size grows rapidly. A much more sensible approach is to have a separate utility which compares documents and prints the new document with changes marked. There is a compare document feature in Word 97 but even then the old and new versions are stored in the revised document. The point I am trying to make is that to implement this feature, desirable for some but not all people, the standard document format has been changed to allow the inclusion of old redundant information.

It is in the interests of Microsoft to change formats in this way because ultimately users get fed up with the problems and upgrade to the latest version, which means more revenue for Microsoft. It is not in the interests of users however and an independent standards organisation could strive to combat it. If the Word 6 format was decreed a standard by an independent body it would be more likely that developers, other than Microsoft, would adhere to it and thus create the competition that the DOJ strives for. In addition, the slowness of change that a standards body instils would be a benefit in trying to maintain a stable situation for users. Microsoft could still introduce their new ideas but there would be a bit of a brake on them. They would be more keen to maintain the standard at the same time as introducing new stuff, which at present they are not.

One point to note is that standards must evolve to cope with new situations and not change too slowly. This is where criticism of standards can be justified. Standards bodies must have the mechanism for changing standards within a reasonable timescale, without the decision making getting bogged down in never ending committees.

Where is the future of document standards? Over the past few years another standard, which caters for some formatting, has grown by default. That is good old HTML. Developed without the help of Microsoft this could be the ASCII of the millennium. Unfortunately HTML has been developed around the concept of screen display and its formatting tags are geared to that screen rather than paper. I assume this is the reason that such fundamental tags as left and right margin are not there. You have to use trickery such as blockquote, or use cascading style sheets, which are an add-on to the basic HTML. I must admit I have started to store some of my documents in HTML rather than Word format. This means they can be read by any browser and also they are usually significantly smaller! I would like to think that HTML would be extended to include paper friendly attributes but I think it unlikely that it will. Despite its usefulness there is a trend against paper from all sides. It costs money to create (in trees), costs money to store, costs money to process and so on. Use a screen and save money. It also removes any accountability because there is no audit trail back to decisions, which is handy for cover up purposes.

We are told that the next stage is XML. XML is a kind of meta language which allows you to define your own tags. This is no doubt going to help extend HTML with the tags that are missing. Even the UK's wonderful government, who likes to think it knows everything about the e-world, has demanded that XML is used as the standard on the NHSNet. Unfortunately, reality is not quite that simple. If we are going to transfer data from one place to another and use it in different systems we all better be using the same tag definitions. The standardisation process comes in defining the tags for a specific type of process, e.g. the collection of medical records.

I hope I have shown in this discussion that standards are essential if we are to work together. Standards, to be useful, need to be fairly stable and do not get altered too rapidly. Microsoft and other developers have a vested interest in rapidly changing de facto standards. Therefore, standards should not be in the hands of developers, but regulated by independent bodies. The independent standard bodies must have a mechanism for changing standards, when required, within a reasonable period of time.