There is a distinction to be made between solutions that accept HTML style content and those that accept real world HTML.
Solutions that accept HTML style or a subset of HTML are fairly small and self contained. However it is important that you have completely control over the HTML you're going to be using so that you can ensure that your content conforms to the capabilities of your solution. The iText XmlTextReader is an example of this.
Other solutions work based on real world HTML and real world browser technologies for proper HTML/CSS/SVG/ etc etc etc support. Our ABCpdf .NET product is an example of this - it includes both a Gecko (FireFox) style layout engine and a Trident (IE - like) engine.
Which you prefer is very much dependent on how much control you have over your source content.