I imagine a DTD page that takes so long to load. Given that it defines entities, you should not disable it , so you should probably not go this route.
Given the internal workings of the wikipedia analyzer (the right mess), I would say that this is a big leap, suggesting that it is going to create well-formed XHTML every time.
Use the HTML Agility Pack for parsing (then you can convert to XmlDocument little easier if necessary, IIRC).
If you really want to go down the XmlDocument route, you can save the local HTML-DTD cache. See this post , this post and this post for details.
source share