(PHp with embedded HTml)
As a programmer, I find myself doing more and more programs that generate output in HTML or XML (in the following, I'll use XML as a generic term for any flavor of HTML and XML itself). None of the languages I use (or know of) are quite adapted to producing this kind of output. For most programming languages, XML is just a stream of arbitrary characters going to the standard output. This made it possible to adopt XML quite fast since all that was required from our programs was to output plain characters, but in doing so we miss a lot. For all a program cares, the XML output is just an arbitrary string, its structure is completely ignored. It shouldn't be so.
XML is a structured language. So are our programs. Both have strict syntax rules. Program translators (compilers or interpreters) take care of checking the syntax of programming languages, but it is up to us to check the syntax of the XML they produce, a most error prone and tiring task. Fortunately, most of us don't even output any XML at all, we either configure ready-made packages, tie together modules and components, use template engines or occasionally use the DOM to manipulate the tree. Nevertheless, at one point or another some part of this whole package does have to produce some sort of XML. Most open source packages such as CMSs, picture galleries, wikis do not rely on any runtime support, neither runtime libraries nor template engines since such dependencies would reduce their portability; they produce XML directly. Yes, lots of XML is still produced directly and there has to be a way to do it better.
So, we have two well structured languages, one is procedural (any flavor
of C, JavaScript, PHP, Perl, etc.), the other descriptive. Their blocks
are quite compatible: to start with, they nest nicely within each other.
If an XML block is contained within an if()
block, it has to
be completely within it, the boundaries of their blocks should not
overlap.
Code is usually embedded into XML via processing instructions like in PHP
<?php ....?>
or other markings, such as
<% ... %>
in ASP or as plain text within an element,
such as <script>
. Nevertheless since most languages
have print
or echo
instructions to output
arbitrary strings to the XML stream, no structure can be forced into such
a program and, to my knowledge, no code beautifier or highlighter can
manage to make sense of this, they can either pretty print the code or the
XML but not both.
Alternatively, other environments such as XML Stylesheets have processing instructions embedded as XML elements, but it is quite cumbersome, just the namespace prefix xsl: or whatever you choose in front of every instruction adds too much typing and the necessary use of attributes to hold instruction arguments, explicitly naming each option, as if you didn't know that the natural argument to a while instruction is a conditional, makes COBOL's verbosity look not so bad.
So, I looked at the problem the other way around, I added XML-specific
instructions into an existing language. I called those instructions
TAG
and ATTRIBUTE
and, since I find no pleasure
in hitting more keys than absolutely required, I decided to shorten them
to a couple of symbols: <
for TAG
and
@
for ATTRIBUTE
. I was working with PHP
at the time and the @
sign was already in use in a manner
which prevented my particular use, thus I switched to the
&
sign. Both <
and
&
are used in PHP, but they are used in ways that a
parser can tell when it is the native usage and when my extended use. And
since I was daydreaming, I opted to use the ?
sign as a
synonym for print
or echo
but with a twist:
since XML tags are already output via the TAG
statement, the
output of ?
will be escaped.