waffle

Hello, OmniWeb. Hello, Omni Group. This is an intervention.

I’ve been using OmniWeb 5 for several years now. When it came out, it contained a bunch of new, unproven ideas. Tabs with thumbnails in a vertical list (that scales from the beginning) and with “loaded but unread” indicators. Site-specific preferences to override obnoxious stylesheets and disable features on some sites. Expanding a form text area to an external window. Workspaces, to persist your sessions and switch to a different set of open tabs and windows. And it was put together in such a way that you could actually hide as much UI as you wanted to just get the web page and the title bar — practically full-screen mode, but without the full screen part. I was a poor student and I didn’t have a credit card, but through some maneuvering, I did actually pay for a web browser (please commence the pointing and laughing) because I thought it was worth it and the ideas were great.

What’s different now from then? a) Lots. b) Nothing.

Nothing is different in that beyond keeping up with first a custom fork of WebCore and then a custom fork of WebKit, adding automatic updates and updated some user agent strings, the OmniWeb team doesn’t seem to have so much as swatted a fly for the past few years. The world has changed around them and although OmniWeb’s aged well with their creative advantage and classic UI, I have a feeling they won’t last a lot longer. OmniWeb is limping.

• Where are my favorites? Thrice, my favorites folder has literally gone missing. I couldn’t reach it from the bookmarks menu or the bookmarks window, but I could hit the number commands (Command+digit for the nth bookmark in the favorites folder; handy for bookmarklets) and activate them. I’ve solved this by removing the Groups.plist file from ~/Library/Application Support/OmniWeb 5/, but the fact that this keeps popping up is unnerving.

• Attack of the tabs. I routinely open a lot of tabs and rely on the “loaded when you were in a different tab” tick mark icon to see what I’ve missed. I don’t mind that it takes a while to scroll along the number of tabs I have open, but I do mind that thanks to persistence, it is sometimes possible to create such a number of tabs (or such heavy tabs) that when you open the browser again, it just starts to beachball, either within a grace period of a few seconds (within which you can manically shut the few tabs within your current reach in the tab drawer and cut your losses) or immediately and you’ll never be able to salvage that session you’ve perhaps been building up all week or all month, saving long things to read for later.

• The source editor, which I used to use a lot more since it is unusually smart in allowing edits and refreshing the associated tab as if you were getting this html directly from the server, somewhere along the line started crapping out and denying saves on opening files with “weird” encoding.

• Finally, the features just aren’t there, thanks to the sliding reality. Where’s my Greasemonkey-like hook, a seemingly obvious extension to the pioneered Site Preferences concept? Where’s my inline Find bar? Where’s the competent ad blocker that can deal with plugins like Flash? Where’s the evolved version of the already great address bar history that Firefox is just starting to catch up to?

OmniWeb team: Why are you not caring about your product, and if you are, why doesn’t it show? Why are you letting people chatter feature requests on your forums without showing some degree of involvement? What’s with not even letting slip that either something is up for the future or that you’re thinking of letting this go — a decision I think many people could sympathize with since maintaining a web browser is a tough racket, and since getting it back to a position similar to where OmniWeb 5 was in comparison to the existing browsers at the time is a steep climb at best?

If you are going to update it, here’s a few places to start:

• Make it a lot more stable, which in essence means that since OmniWeb doesn’t crash a lot of its own volition, it needs to be resistant to plugins crashing. It’s basic math that when I house 107 tabs in it, as I did the other day, the risk will be higher that Flash Player will jump out of a window on some odd tab and take the whole app with it.

  Yes, I am advocating moving to tabs and/or plugins out-of-process, like Google Chrome and IE8. Grumble for a while and then go ahead and do it. Or think up a better solution. Or prove to me that this kind of crash is actually acceptable.

• If at all possible, drop the WebKit fork. (I may just have negated my previous pointer, which should take precedence. But the advice stands — if it’s possible, you should drop it.)

• Modernize the UI. This basically encompasses three steps as far as I’m concerned. Move the tab drawer to be a collapsible same-window panel; do the inline Find bar thing and assume that when I’m typing text into the address bar without a prefix, I want to search with the default engine. Google Chrome has proven to me that this can be done in a sane manner (I was highly skeptical), and I will no longer put up with anything else.

• Allow for editing workspaces and lazy-load tabs by default. This solves the death-of-a-thousand-tabs issue and improves startup performance.

• Allow for new kinds of plugins. What I’m missing today is a programmable browser. I mean something beyond Firefox, which is modularly built and where you can plug into everything. I mean something with a fixed set of well-defined extension points that are all useful. Imagine being able to hook into a tab live and get at the already loaded media, the page the way it looks right now… that kind of runtime metadata. Adblock Pro has proven that given the sufficient hooks, your community will write your ad blocker for you, and this would be a great place to start.

And, you know, if you really are thinking of calling it quits… make it open source. If I close my eyes and think of England, I could see an OmniWeb rejuvenation effort take off.

John Siracusa once praised OmniWeb with these words:

Usually this level of functionality can only be found in the geekiest of open source web browsers, if it can be found at all. Finding it in a proper Mac OS X application from a respected developer with a proven track record is like finding a perfect 1/10,000th scale replica of the Eiffel Tower in a box of crackerjacks. Then the tower transforms into a tiny robot and makes you lunch.

I felt that way when I got my hands on this browser for the first time. I’m not inspired enough to actually continue the metaphor, but suffice it to say that the feeling’s waning. And for an application that was once so great, I don’t want it to end that way.

Update: Ken Case, CEO and founder of the Omni Group, has read this post and commented on it. I’m pleased that top men (top!) at the Omni Group are keeping informed and acknowledging OmniWeb’s four year stint out of focus, and I’m looking forward to seeing what they’ll think up when OmniWeb 6 does come into being. I certainly appreciate the uphill climb of creating a web browser that’s not only on par with the others but that has enough of an edge that you’d pay for it.