Fronteers considered harmful
============================

Monday 22 March 2010 20:54

Addendum: By "fronteers" I don't mean the dutch
professional organization "Fronteers", but more
generic people who very passionately try to make
the impossible possible by pushing the envelope of
web standards in a pioneering, front running,
frontenders way of using HTML, CSS and JavaScript
in their struggle with the browsers. All I remember
is ppk inventing this label that covers this area
exactly how I see it, but apparently around
Amsterdam the term has since gained more popularity
as the name of that organization. (Addendum 2: If I
want to refer to that organization I'll use the
term "bunch of unprofessional HTML monkeys"
instead.)

The problem with some front end developers is that
they believe everything that is related to HTML,
CSS and JavaScript should be aimed at making their
job easier, because it's the hardest job in the
world. They see standards as solely written for
their line of work, totally ignoring the fact that
there are actually real developers who have to
implement those standards into a browser first, and
not every situation the standards have to cover has
the convenience of upgrading to the latest mozilla
release candidate to fix some issues.

Consider the following scenario. Some manufacturer
of a piece of hardware wants to have a fancy user
interface on the thing. The piece of hardware could
be a television set, a fridge, industrial cleaning
equipment, whatever. The manufacturer can slap a
little LCD touch screen on it and put a cheap board
in it that provides the user interface. In a
perfect world, shouldn't that user interface be
build with open standards like HTML, CSS and
JavaScript? So the manufacturer sets out to embed a
browser in the thing. The manufacturer can buy the
required software components that are already
available and put them in. But because the embedded
browser is limited to that specific device it could
take advantage of features specific to that device.
The embedded browser can be extended to handle CSS
that isn't standard so the developers of the user
interface have a few extra properties to easily
implement the specific features.

The CSS standard takes care of this by allowing
vendor prefixes. Without vendor prefixes there is a
chance that a newer version of the standard would
introduce a property that has the same name as
something the vendor has invented, causing a
conflict and confusion. So vendor prefixes are a
good thing.

We can't get away with generic -draft- or -beta-
prefixes because our manufacturer could then
implement their features as a -beta- property, and
the upstream provider of the CSS engine for the
embedded browser could suddenly also introduce that
same -beta- property for something else, causing
the same conflicting situation that could occur
without vendor prefixes. So the upstream provider
should use something like -upstreamprovider- and
the manufacturer something like -manufacturer- as
vendor prefix. And when the manufacturer then sells
the code of it's embedded browser to another
manufacturer they can do the same thing and every
update from the upstream provider passes the
workflow without problems.

The only problem is that currently a large part of
the browsers used for surfing the web on desktop
and mobile devices is based on open source engines,
which makes it possible for web developers to find
out what all their vendor specific features are and
what they do. Adopting all these vendor specific
features to do cool things before they become a
proper standard - and the competition does the same
cool things - web developers walk straight into the
trap that was fueling the browser wars in the 90s.
Vendor prefixes aren't defined as being latent
standards and it's just wishful thinking to see
dropping them as a solution for easier web
development, or speeding up some standarization
process. But apparently web developers have to
waste the time they gained from dropping support
for Internet Explorer 6 by setting up the next self
induced mess that makes their job so hard.

by Roland van Ipenburg
http://www.xs4all.nl/~ipenburg/blog/posts/work/2010/03/22/fronteers-considered-harmful/