DOM Level 3 a W3C Recommendation

It’s taken a long time, far longer than we anti­cip­ated when we star­ted. The W3C DOM Level 3 Core and Load and Save mod­ules are, as of today, W3C Recommendations.

The DOM WG figured out in the very first meet­ing, back in March 1997, that we would have three levels (each level builds on the pre­ced­ing level; it’s a way of mod­u­lar­iz­ing the spe­cific­a­tion that is rel­at­ively easy to under­stand). I don’t think any of us expec­ted it to take 7 years!

I chaired the DOM WG until Novem­ber 2001, enough time to get Level 2 out as a set of Recom­mend­a­tions. Phil­ippe Le Hégaret and Ray Whit­mer chaired the group after I left.

The early days of the DOM WG were a polit­ic­al roller­coast­er, of course. These were the days of the Netscape/Microsoft browser wars; we were hav­ing a DOM WG meet­ing at Nets­cape at the time that someone placed the IE “e” on top of Nets­cape’s sign at their cor­por­ate headquar­ters. The DOM sup­port in the browsers was one of the battle­grounds but to give every­one on the WG cred­it, most people did genu­inely try to find tech­nic­al solu­tions to tech­nic­al prob­lems with as little cor­por­ate polit­ics as pos­sible. To help with this, sev­er­al con­fid­en­ti­al­ity policies were insti­tuted at W3C to pro­tect mem­bers of the WG from journ­al­ists try­ing to find a story (yes, stor­ies exis­ted, but I wanted to get some work done, and pub­lish­ing them would not have helped). As Chair of the WG, I could talk to journ­al­ists (people oth­er than the Chair and the W3C staff con­tact were not meant to talk on behalf of the WG, although of course if they wished they could talk on behalf of their com­pany as long as they stuck to things that the WG had agreed could be made pub­lic). I remem­ber one occa­sion when a journ­al­ist called up, ask­ing me to tell them of occa­sions when Nets­cape had “won” a tech­nic­al dis­cus­sion, and Microsoft had “lost”. Sigh. Obvi­ously I declined. 

The biggest splits actu­ally came between those who were “HTML-cent­ric” and those who were “XML-cent­ric” � the “array” style of walk­ing the doc­u­ment and the “tree” style of walk­ing the doc­u­ment came about because neither side could con­vince the oth­er that their trus­ted meth­ods for find­ing the ele­ment they wanted were not needed. So the DOM Level 1 ended up with both ways of look­ing at an XML or HTML document. 

Polit­ics still gets in the way of stand­ards work; I hope one effect of the Sun/Microsoft pact will be to allow the tech­nic­al people on the tech­nic­al com­mit­tees more free­dom to come up with the best tech­nic­al solu­tion to tech­nic­al problems.