[Simon]

OK, we have just got Safari and with it the WebCore classes. Apple has it's own browser and it wors already pretty good for a beta. It will be on every OS X Mac sold.

Some people have suggested that WebCore could be used as the engine behind Sherlock and HelpViewer. Sure this could make life a lot easier for Apple and I'm sure both would probably perform better, but still, why?

Sherlock is dog slow and although it makes some stuff much easier than "by hand in the web" it is left standing around somewhat lost. You don't need it for finding files anymore, since we can search those through the Finder directly. Today, Sherlock remains something for web users like the "Services" menu is to Finder users. HelpViewer on the other hand might be a nice idea, but the interface lacks a lot of browser navigation flexibility and it's not exactly a speed daemon either.

So what's the new approach? Imagine Apple just folds Sherlock and HelpViewer into Safari like Safari's bookmarks. A small icon (one for Sherlock, one for HelpViewer) in the bookmark bar (or menu item - might as well make it an option) which upon clicking opens the "like-iTunes" interface Safari also has for its bookmarks. In the list on the left side you get the Sherlock channels or the list of application help available. You click on an item and on the right (in the "browser" window) you get the Sherlock channel interface or the HelpViewer contents.

I suppose you could also implement direct access to these features like it has been since OS X came out: You click on help, Safari opens up in its HelpViewer "mode" and shows you the help contents for the specific app you were in. For Sherlock just fire up Safari and go to the Sherlock panel. Safari sure is up and running much more quickly than Sherlock is.

Apple saves a lot of work by not writing a viewer engine for three apps - only the WebCore engine from Safari is needed.
Developers can write Sherlock channels that every Safari user can use. Double click the channel file, it opens in Safari, goes to the Sherlock panel and at the same time copies the file to the right directory in the user's ~/Library folder (much like iTunes if I double-click an MP3 file). Developers also can use simple HTML with no limitations as help files. Since HelpViewer's engine is Safari every valid HTML works. Much better than today's HelpViewer which is some kind of hybrid between HTML and some Apple limitation.
Finally, users get the speed but also the navigation flexibility Safari offers. I hate the HelpViewer's navigation - it looks like an HTML browser, but it feels and works much less like any browser I know. I'd love it to be just as good as Safari.

What do you guys think? Do you believe Apple has already been looking into something like this?

[undotwa]

Interesting...

Sherlock could become a framework 'netservices.framework' which any app which wants to call it can (and of course Safari would call to access net services.)

Safari replacing Help Viewer is pretty darn logical, and practical.

Just like 'Address Book' has become a framework which any app can call (Mail, iChat, Safari...).

Maybe like the Bookmarks how it opens over the current window, and a similiar design with the left showing subscribed Net Services (configurable in any front end...) and the right pane the active net service.

How's that?

[Simon]

Originally posted by undotwa:
How's that?

Right on.

[willed]

Sounds like a plan. Let's do it!

[edddeduck]

I don't like the idea as Safari is a web-browser not a joat (Jack Of All Trades).

As they say jack of all trades, master of none.

If you need help you don't want to have to open Safari you just want help. Help will be fast now if they use the safari webcore.

As for sherlock I can see the sence but if you look at sherlock it's the interface that makes it usefull. And the crap html engine that makes it slow.

Also for at least 50% of the people it is useless as we don't live in the US.

Finally I don't see it happening that Safari will merge with Sherlock because that would all alot of bloat/size to the app and Apple (and me) want a fast lean standards compliant browser. We don't want another Netscape.

But It would be cool if they made sherlock available in services or links between them but for speed AND ease of development they should be 2 seperate apps just like they are 2 dev teams in real life.

Merging them would mean development would slow down as you will have people learning each others code and they both have totally different goals one team search engine one web browser.

Finally Apple would have the "make sherlock opensource" crowd on their back.

(Sorry I didn't mean to put a downer on an upbeat thread but I had to say it.)

Cheers Edwin

[Guy Incognito]

Edddeduck is right...keep apps seperated.

It doesn't surprise me undotwa loves this stupid idea though.

[Simon]

Originally posted by Guy Incognito:
Edddeduck is right...keep apps seperated.

That's exactly the point. HelpViewer should not be an app. It's just a system-wide help browser. One that relies heavily on HTML. There's no reason to have two programs doing the same thing, only that one does it right and the other does it half-hearted. Sherlock is its own app and that makes some sense. However, you will have to agree that all it does is heavily web-orientated. It's not just something than uses TCP/IP to exchange content, but it does HTML rendering of HTML content. Why confuse newbies and bloat the app folder if it's just about web content?

It doesn't surprise me undotwa loves this stupid idea though.

How nice of you. What great effort you must have put into that thought.