Re: [xml-dev] Many different syntaxes in XML - is that goodlanguage desi

[Shaun McCance <shaunm@gnome.org>]

On Mon, 2022-03-07 at 12:14 +0000, Roger L Costello wrote:
> Hi Folks,
> 
> There are many different syntaxes in XML:
> 
> The comment syntax:  <!-- ... -->
> 
> The PI syntax:  <? ... ?>
> 
> The CDATA section syntax:  <![CDATA[ ... ]>
> 
> The DOCTYPE syntax:  <!DOCTYPE ... >
> 
> The entity syntaxes:  &amp;  and &#x020; and &#032;
> 
> The namespace syntax:  xmlns:foo="..." and <foo:bar>...</foo:bar>
> 
> Attribute/value syntax:  x="blah"
> 
> The start-tag/end-tag syntax:
> 
> <baz>
>    <widget>...</widget>
> </baz>
> 
> Phew! That is a lot of different syntaxes.
> 
> Is that good language design to have so many different syntaxes?
> 
> Dealing with lots of different syntaxes is hard. For example, I never
> remember the syntax for CDATA sections (in fact, I had to look it up
> just now).
> 
> Isn't it better language design to have a small number of syntaxes?

Counter-point: one of my favorite features of XML compared to a lot of
other text formats is that it has very few characters that are actually
syntactically relevant. When I write in many "lightweight" formats, I'm
second-guessing whether some character I typed is going to trigger some
parser feature. When I'm writing text in XML, the only two characters I
have to worry about are < and &.

If I were designed XML from scratch today, there are things I'd do
differently, including that CDATA syntax. But one thing I wouldn't
change is the predictability of characters.

--
Shaun