"Eric J. Schwarzenbach" <[EMAIL PROTECTED]> wrote on 12/27/2006 02:03:28 PM:
> Isn't a billion characters far more that you could ever load into a DOM > in any practical environment? In general yes. However, if you know in advance which regions of the document actually require random access and which parts of the document you will never visit it may be possible to keep the memory usage low by using a filter [1]. > The question I find more interesting that of the feasibility of VTD as > general DOM / Xerces competitor, is could some of the underlying > techniques [1] they use in processing the XML be used to make a more > performant and scalable DOM parser? In their FAQ they suggest no [2], > but they seem to assume that the whole DOM object structure need be > created and not just parts of it lazily as they are actually used. I understand Apache AXIOM [2] (which is an XML object model used by Axis2) will lazily contstuct its object tree by pulling document events from the parser as needed. There's no reason you couldn't do the same (or even better) with DOM. > Eric [1] http://www.w3.org/TR/2004/REC-DOM-Level-3-LS-20040407/load-save.html#LS-LSParserFilter [2] http://ws.apache.org/commons/axiom/index.html Michael Glavassevich XML Parser Development IBM Toronto Lab E-mail: [EMAIL PROTECTED] E-mail: [EMAIL PROTECTED] --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
