Stefan Seefeld wrote:

Phil Endecott wrote:

...

Does anyone have any experience of how little overhead could be involved in going from

e.attributes["foo"]="blah"; to e.set_attribute("foo","blah"); ?

That is pure syntactic sugar. Yes, this can be done using some proxy classes. Why should we worry about such details at this point ?

The syntax is important. What are the requirements from an XML library? I would argue that an important requirement is an easy-to-learn syntax. Many potential users will be familiar with other libraries and if they can apply that experience to a new library it will make it quicker and easier to use. Hence my original argument that there are two justifiable APIs: one that is familiar to people who know the DOM syntax, and one that is familiar to people who know the STL syntax. Of course there are other potential requirements. Performance is one, but my (humble) opinion is that a _Boost_ XML library would focus more towards an STL-like API and less towards performance and "thin-ness". I'm sorry that my comments are sounding negative. I have used xmlwrapp, which is similar to your library, and it did a good job. I have no doubt that your library would be useful. But you haven't convinced me that your API, or the approach of wrapping libxml2, is the right style for Boost specificially. Cheers, Phil.