[Cryptography] encoding formats should not be committee'ised

[Jerry Leichter]

On Oct 3, 2013, at 7:33 PM, Phillip Hallam-Baker <hallam at gmail.com> wrote:
> XML was not intended to be easy to read, it was designed to be less painful to work with than SGML, that is all....
More to the point, it was designed to be a *markup* format.  The markup is metadata describing various semantic attributes of the data.  If you mark up a document, typically almost all the bytes are data, not metadata!

TeX and the <x>-roff's are markup formats, though at a low level.  LaTeX moves to a higher level.  The markup commands in TeX or <s>roff or LaTeX documents are typically a couple of percent of the entire file.  You can typically read the content, simply ignoring the markup, with little trouble.  In fact, there are programs around at least for TeX and LaTeX that strip out all the markup so that you can read the content as "just plain text".  You can typically get the gist with little trouble.

If you look at what XML actually ended up being used for, in many cases nearly the entire damn "document" is ... "markup"!  The "data being marked up" becomes essentially vestigial.  Strip out the XML and nothing is left.  In and of itself, there may be nothing wrong with this.  But it's why I object to the use of "markup language" to describe many contemporary uses of XML.  It leads you to think you're getting something very different from what you actually do get.  (The XML world has a habit of using words in unexpected ways.  I had the damnedest time understanding much of the writing emanating from this world until I realized that when the XML world say "semantics", you should read it as "syntax".  That key unlocks many otherwise-mysterious statements.)

                                                        -- Jerry