This has been a really good conversation. I’ve also gotten some excellent off line emails articulating the issues. As I have been reading the feedback, noodling over it and discussing it with the help team, I’ve come the the following summary conclusion:
We set out to create an improved help platform that is open, standards based and extensible. The primary benefits we were after were a good online experience, a very customizable off line experience without redundant content authoring, the ability for others to contribute to the offline content, a simple default UI and a customizable presentation experience (the new help SDK makes it easy to build alternate help clients). In doing so, we took a step backwards on some valuable help features. Over time, we will add back many of those features and refine the help system to provide an even more seamless help experience. The difference is that it will be built on this open and extensible platform that enables innovations not possible before.
I do believe we talked to a lot of customers about the experience we were building but I also believe that what I’ll call the “power help user” may have been under-represented. Much of the feedback has helped me understand the kind of experience that the power help user is looking for. I believe the help team has internalized much of the feedback and has ideas on how to continue to improve the experience.
There were a number of questions/comments on my last post on help that I think were misconceptions about the new help system (in part based on behvior of the last).
1) The new help system is asynchronous. Your IDE does not lock up while help is launching. Also, it’s generally pretty fast, by the way, so even if it weren’t asynchronous, most of the time you wouldn’t notice.
2) It does support offline. I realize the natural assumption of using a browser to host the content is that it is accessing the internet. While it can, it can also serve the content from the local machine.
3) I think I really get now why some of you are so wed to the index, I can see how it can be very helpful.
4) I totally agree that relying exclusively on the browser for the window management probably isn’t going to be a great experience and we will need to provide an alternate one.
I’m not going to respond to each and every comment/question (there were quite a few), but I’m grateful for the feedback and if you think I’ve missed something important or misinterpreted something, please let me know.