Maybe now they can make a better web browser, hopefully. I would have chosen to design some things a bit differently, including a better separation of parts tha

Maybe now they can make a better web browser, hopefully. I would have chosen to design some things a bit differently, including a better separation of parts that can avoid binding too closely (to make it more possible to customize it in programs that use these libraries), if it does not already do (I do not know how closely they already are; I haven't read all of it). I would also put the protocol interface separate from the implementation (so that other protocols can more easily be added). (I also would have preferred C instead of C++, but C++ might also do OK, anyways.) Also should be helpful that independent programs can be written which combine them and use them in their own ways, e.g. to change the connections between components, to deliberately exclude some features, and/or to add new protocols and URI schemes and file formats and JavaScript functions and HTML commands and CSS commands (I have idea how to do by "meta-CSS", which is a feature to be used purely by the end user (document authors cannot use it)) and character encodings and audio/video codecs and other features, and/or changing some features. (I would hope to be able to write such extensions/programs in C, instead of having to use JavaScript.) I also would have designed that it avoids converting to/from Unicode if possible, preferring to work in the text encoding directly, and using fonts of those encodings directly, when possible, and using Unicode as a fallback when needed or if there aren't fonts, etc available for the specific encoding already used. (This is to work around many problematic features of Unicode.) User options are also needed. For example, there should be an option to disable MIME sniffing if the user does not want it. (One way to support this and other options is to implement a "header overriding" menu that the end user can configure (both request headers and response headers); this can be used to implement many options that the end user can configure, including (but not necessarily limited to): DNT, language, cookies, HSTS, etc.) Also, a better implementation would not try to hide things from the user or believe they know better than what the user has specified. I also think that some parts of the W3C specification aren't so good and need to work around that somehow, in some cases. Nevertheless it seem a good idea that they can make a new one in order hopefully is better than the other one (which, in time, we may be able to see).