|
It is noteworthy how unpopular approaches can get new life after a period of dormancy; just observe how much fellow commenter Mr. Bullard sounds like me circa 2002, dusted off for a fresh wave of saucy weblog snarking :)
Your article reinforces my own tack. I will freely admit that the ways I follow progress in software is based on Larry Wall's laziness virtue, but that doesn't make them bad. I keep reading, but as I assume all developers do to a degree I assign risk to technologies in flux because I don't like to waste time learning the syntactical details and subtleties of some paradigm that won't have lasting relevance. For the record, neither working with RSS 1.0 nor naming my fledgling software project in the article title (http://www.onjava.com/pub/a/onjava/2003/07/16/ldaphttp.html) were my idea.
Since my earliest professional days, Microsoft was a non-starter (remember Windows 3.1?). I skipped Motif at the advent of the web, I saw SGML as powerful but too complicated, and I yearned to do far less tcl programming than I was eventually forced to. Instead I invested most heavily in HTML, HTTP, Perl, and Java, which have all proven to be keepers.
I knew XML would be a winner, but the many standards that immediately followed were nascent. I think SOAP misses the point completely even if it's become the boat. XHTML is appropriate, though parsing it honestly doesn't do much for screen scraping. Many web service languages translate to spaghetti for the 21st century enterprise even if the burden of understanding them will be split more evenly between the computers and programmers. I'll add that you weren't the only one holding out for a better XML/CSS union, and I'm not convinced it's time to give up yet considering how weakly web browsers support XML in general.
Then there's AJAX. I systematically avoided serious reliance on Javascript because of variablilities in its implementation, strange constructs in its coding, and uncertainty in its future. Alas, the problem has always been there: no other broadly supported way to make web interfaces rich. AJAX doesn't prove that instict wrong, it proves I was right because now I can potentially spend a shorter amount of time learning a consistent XML language for solid interfaces instead of wasting time writing and testing questionable homebrew Javascript routines.
The semantic web has been an idea in circulation for a while, and it seems to describe what we're moving towards even if the specifics of the vision haven't been implemented in one swipe. With the momentum and ubiquity of LDAP and HTTP protocol use on our networks, it occurred to me that their synergy could give us something close. That's my own cross, perhaps, but judge for yourself (http://www.mentata.com/ldaphttp/examples/mesh/) if combining them, couching the calls in REST, and making the data simultaneously available directly from the directory (sorry, not yet demonstrated outside my private network) makes any sense. Even if you don't browse any of the thousands of resulting pages, Google will. Note also that even a semantic web benefits from identity knowledge.
In the meantime, while I agree that Web 2.0 is more a brand name than a well-articulated vector, it says in a phrase where we want to go. Good luck filling in those blanks, and thanks for the thoughtful essay.
|
Previous Message
