I know I'll probably be a voice of 1 but I have a serious hatred for most of markup languages. HTML is a fun little thing kind of like "Reveal codes" was fun in Word Perfect and it makes for some simple formatting. XML seems to be used primarily for data transfer and I hate it for multiple reasons. Even XML's so called "attributes" easily more than double or triple the size of the file (I've seen files with ratios in the hundreds). For the sake of argument, I'll just say that it doubles the size of most tabular data. That doubles the size for storage on disk, usually more than doubles the size in memory because of the cruddy "edge" tables that get built behind the scenes to parse XML, doubles the transmission time because there are at least double the number of bytes just from the tags never mind always being character based instead of 1, 2, or 4 byte integer based for most numbers, doubles the size of the cooling systems I need because everything has to work at least twice as hard to parse the data, etc, etc, etc. Heh... and who in the world actually thinks it's a good thing to depend on nested non-unique tags to build a hierarchy with? And then people clog up my databases with this crap because the "need to use the XML all at once" from the front-end. Yeah... the real reason they need it that way is because a lot of them don't know how to parse it and normalize it before dropping it into a database. Don't even get me started on EDI! :-P
Phil wrote an article a while back where he also (IMHO) appeared to be a bit disgusted with the byte-bloat caused by markup languages for the transmission of data. He suggested that someone come up with a better way and I was going to jump at the chance except for one thing... one of the best ways to do all that was invented scores of years ago. Lookup ASCII characters 28 through 31 and see just how simple and fast data transmission used to be. If we reverted to such a thing for the transmission of "unformatted" (i.e. not for display purposes) with a little common sense of transmitting byte values for numbers instead of characters, most of us would never have a "performance issue" with our networks, storage systems, etc. Instead of having bottlenecks, our data would be like a BB shot from a rail-cannon through the Holland tunnel instead of like trying to drive a pig truck through a Coke bottle.
We just don't need the overhead of languages like XML to transmit even some of the most complicated data. I won't get into what a travesty I think it is to transmit tagged hierarchical data to an RDBMS even wheen it seems like a good idea. And, no, we don't need formatting tags (even when CSS is available) on every bloody element of data even for the front end. Assign a "default" and only include formatting for exceptions. I know that's the way that a lot of good people try to do it but, apparently, none of the people sending me XML data are "good" people. :-D
And before you folks blocking the back door of this club go postal on me about human readability, isn't THAT what the presentation layer is for???
is pronounced ree-bar and is a Modenism for R
First step towards the paradigm shift of writing Set Based code: Stop thinking about what you want to do to a row... think, instead, of what you want to do to a column.
If you think its expensive to hire a professional to do the job, wait until you hire an amateur. -- Red Adair
How to post code problemsHow to post performance problemsForum FAQs