Add new functionality now or later?


One thing we had to think about in TC45 when dealing with the national body comments was how to deal with requests for additional features or functionality. It’s important to make sure at this stage to only fix issues in existing functionalities and that you don’t put things into a standard that haven’t been properly vetted.

For example, while I was surprised that ODF didn’t have formulas in the first version of their specification even though spreadsheet formulas had already been so widely used in the industry, it wasn’t as big of a deal that they don’t have rich ink (handwriting) functionality as that’s a bit newer in terms of office document integration and could probably use a bit more thinking in the ODF camp.

There’s a great quote from Charles Antony Richard Hoare in his book “Essays in computing science”:

When any new language design project is nearing completion, there is always a mad rush to get new features added before standardization. The rush is mad indeed, because it leads into a trap from which there is no escape. A feature which is omitted can always be added later, when its design and its implications are well understood.

This is why in TC45, we made changes where we felt we had the appropriate fixes and extensions to existing functionalities that were requested by National Bodies. We did this where we felt we had adequate time to think through the design. In other cases though, we had to push back and say that the suggestion was valid, but it was more appropriate to handle in future maintenance. Of course the final decision has to be made by the national bodies during the BRM.

-Brian

Comments (4)

  1. Karellen says:

    Wow. Interesting that someone from MS would bring  up the spreadsheet formulae thing again, especially when making the point "that you don’t put things into a standard that haven’t been properly vetted."

    The first time this was brought up by MS, they had also not yet defined formulae in OOXML. Then, they themselves rushed their current implementation into OOXML as it was, clearly having not vetted it, as it contained a number of omissions (some of which I’m led to believe are now fixed, such as failing to specify units for trig functions, some of which might not be) and a number of clear problems do do with internationalisation and localisation; a large number of formulae make US or Western (or sometimes Eastern, and are similarly non-portable to the west) assumptions that make those formulae unsuitable for an international standard.

    Maybe if MS had taken the time over spreadsheet formulae to vet the work properly (as the OpenFormula committee are doing for the formula work they will be submitting for ODF1.2) then OOXML would have been in better shape for submission to a standards body.

    In fact, given the state of OOXML as originally given to TC45, it appears that taking their time and properly vetting things before they went in to the standard was never a concern for Microsoft before. Maybe if it had been, there wouldn’t have been 3,000+ comments to deal with in the first place.

    Of course, I’m not complaining if MS are changing their ways to be more thorough now. It just seems particularly ironic and/or hypocritical for MS to bring this particular charge against people who are trying to fix deficiencies, given that those deficiencies were apparently caused by MS operating in exactly the same way – not spending "adequate time to think through the design" – to begin with!

  2. jones206@hotmail.com says:

    Karellen,

    I don’t know what you mean about the formula specification  not existing when we started to point out this flaw in ODF. I think Tim Bray from Sun was the first one to point out this shortcoming in ODF, and I followed.

    The formula’s in Excel had been well defined for quite some time. OpenXML was submitted in draft form in December of 2005, and within a few short months the first draft of the formula piece was available for public preview.

    It’s well defined and available for anyone to implement.

    Compare this to the SVG and XForms support in ODF where noone has been able to claim the fully implement it as it’s so vaguely defined.

    -Brian

  3. yoonkit says:

    This "Formula Specification" in MSOOXML has been claimed to be such an amazing addition, I had a look at it several months back now.

    Its nothing more than a dump of the Microsoft Excel help files, warts and all.

    Please read this post: "Mathematically Incorrect"

    http://www.openmalaysiablog.com/2007/07/mathematically-.html

    It compares the Formula Spec issue in MSOOXML and in ODF, and it clearly shows how one is superior than the other when multiple vendors views are addressed to build a consensus within a standard.

    yk.

  4. gerd says:

    "One thing we had to think about in TC45 when dealing with the national body comments was how to deal with requests for additional features or functionality. It’s important to make sure at this stage to only fix issues in existing functionalities and that you don’t put things into a standard that haven’t been properly vetted."

    Fast-track is no process for standard development. It was a mistake of the ISO members to let you enter that procedure.