Comments on: The Problem With Atom http://www.symphonious.net/2008/09/02/the-problem-with-atom/ Living in a state of accord. Fri, 12 Mar 2010 07:41:05 +0000 http://wordpress.org/?v=2.9.2 hourly 1 By: Adrian Sutton http://www.symphonious.net/2008/09/02/the-problem-with-atom/comment-page-1/#comment-173299 Adrian Sutton Wed, 03 Sep 2008 09:07:23 +0000 http://www.symphonious.net/?p=949#comment-173299 Having used both WordPress and WCM for the same site, there are trade offs both ways. WCM doesn't have feeds baked in, but WordPress lacks workflow. WCM is difficult to set up but extremely flexible, WordPress is fast to get up and running but then took as much effort as WCM in the end to get some of the elements of the site and they were far less maintainable. I'd have the same horror story for you about getting dynamic side bar menus in WordPress or the plugins listing on the front page, or the changelog on the early access stuff. I've essentially had to develop plugins and custom code for both WCM and WordPress. Drupal is probably a good middle ground and is likely to be the ideal type of system for LiveWorks! but we do a huge amount of partnering with IBM and want to dog food our integration with their products. Besides which, I'll sound a lot more intelligent talking to WCM clients now that I know how the whole system works in the real world, not just the editor bit of it. Having used both WordPress and WCM for the same site, there are trade offs both ways. WCM doesn’t have feeds baked in, but WordPress lacks workflow. WCM is difficult to set up but extremely flexible, WordPress is fast to get up and running but then took as much effort as WCM in the end to get some of the elements of the site and they were far less maintainable. I’d have the same horror story for you about getting dynamic side bar menus in WordPress or the plugins listing on the front page, or the changelog on the early access stuff. I’ve essentially had to develop plugins and custom code for both WCM and WordPress.

Drupal is probably a good middle ground and is likely to be the ideal type of system for LiveWorks! but we do a huge amount of partnering with IBM and want to dog food our integration with their products. Besides which, I’ll sound a lot more intelligent talking to WCM clients now that I know how the whole system works in the real world, not just the editor bit of it.

]]> By: Asbjørn Ulsberg http://www.symphonious.net/2008/09/02/the-problem-with-atom/comment-page-1/#comment-173298 Asbjørn Ulsberg Wed, 03 Sep 2008 08:24:43 +0000 http://www.symphonious.net/?p=949#comment-173298 You make me feel like the happiest man alive for not having to poke around with IWWCM. Sounds like a pretty horrifying CMS to me. It sure seems enterprisey, though. If that's a positive or negative label is within the eye of the beholder I guess. Personally, I hate everything purporting itself as "Enterprise Software". Which is why I'd choose WordPress (or Drupal) any day over multi-million dollar solutions. At least they're FOSS, so if they don't match my needs exactly, I can tweak and bend them in any direction and way I want. You make me feel like the happiest man alive for not having to poke around with IWWCM. Sounds like a pretty horrifying CMS to me. It sure seems enterprisey, though. If that’s a positive or negative label is within the eye of the beholder I guess. Personally, I hate everything purporting itself as “Enterprise Software”. Which is why I’d choose WordPress (or Drupal) any day over multi-million dollar solutions. At least they’re FOSS, so if they don’t match my needs exactly, I can tweak and bend them in any direction and way I want.

]]> By: Adrian Sutton http://www.symphonious.net/2008/09/02/the-problem-with-atom/comment-page-1/#comment-173296 Adrian Sutton Tue, 02 Sep 2008 14:32:46 +0000 http://www.symphonious.net/?p=949#comment-173296 You could post process it using a reverse proxy server type set etc, but IWWCM itself literally has 3 fields: start, item and end. You can only refer to specifically named content items in the start and end sections and the item section is repeated for every content item that gets output. You can generate the updated element for each item because in that field you have access to the current item, but you don't have access to the first item in the start field. You also can't create the updated element in the item field because then every content item would generate it, giving you something like: <updated><item>...</item><updated><item>...</item> I think you could do it by adding a second menu that only ever output 1 item (the most recent) and you could include that menu by name in the start section. With the way IWWCM iterates through content to find items to include in menus though that would be almost equivalent to generating the entire feed twice for each request. It's also way more complex that you'd want just so that you can have an updated element. Oh and there's no XML structure in IWWCM - you're just outputting text that happens to be XML when it all comes out. IWWCM doesn't know it's generating XML - and in fact it declares the content type as text/html which is rather unfortunate. You could post process it using a reverse proxy server type set etc, but IWWCM itself literally has 3 fields: start, item and end. You can only refer to specifically named content items in the start and end sections and the item section is repeated for every content item that gets output. You can generate the updated element for each item because in that field you have access to the current item, but you don’t have access to the first item in the start field.

You also can’t create the updated element in the item field because then every content item would generate it, giving you something like: ……

I think you could do it by adding a second menu that only ever output 1 item (the most recent) and you could include that menu by name in the start section. With the way IWWCM iterates through content to find items to include in menus though that would be almost equivalent to generating the entire feed twice for each request. It’s also way more complex that you’d want just so that you can have an updated element.

Oh and there’s no XML structure in IWWCM – you’re just outputting text that happens to be XML when it all comes out. IWWCM doesn’t know it’s generating XML – and in fact it declares the content type as text/html which is rather unfortunate.

]]> By: Asbjørn Ulsberg http://www.symphonious.net/2008/09/02/the-problem-with-atom/comment-page-1/#comment-173295 Asbjørn Ulsberg Tue, 02 Sep 2008 13:56:17 +0000 http://www.symphonious.net/?p=949#comment-173295 How can the IWWCM menu structure not handle the "updated" element? Can't you drive it through XSLT or otherwise manipulate the XML structure somehow? The "updated" element is there, as you write yourself, for easy consumption. How can the IWWCM menu structure not handle the “updated” element? Can’t you drive it through XSLT or otherwise manipulate the XML structure somehow? The “updated” element is there, as you write yourself, for easy consumption.

]]> By: Adrian Sutton http://www.symphonious.net/2008/09/02/the-problem-with-atom/comment-page-1/#comment-173289 Adrian Sutton Mon, 01 Sep 2008 21:56:45 +0000 http://www.symphonious.net/?p=949#comment-173289 Yeah time.now isn't available either but that's good to know for most other situations. I'm sure I could do some more Apache regex foo as well but RSS 2 is simpler and works just as well for this use case. IWWCM has a whole swag of caching stuff that I haven't delved into yet but by default it recomputes the feed every time. The feed IWWCM generates is only fed into FeedBurner and FeedBurner is smart enough to do it's own caching etc so the impact is minimal. The bigger problem is the lack of caching on a number of resources like JavaScript files etc - at least the theme images are served directly from Apache so are fairly cacheable. Overall though the site is a pretty spectacular example of bad caching practices - just like it was when it was running on WordPress. Yeah time.now isn’t available either but that’s good to know for most other situations. I’m sure I could do some more Apache regex foo as well but RSS 2 is simpler and works just as well for this use case.

IWWCM has a whole swag of caching stuff that I haven’t delved into yet but by default it recomputes the feed every time. The feed IWWCM generates is only fed into FeedBurner and FeedBurner is smart enough to do it’s own caching etc so the impact is minimal. The bigger problem is the lack of caching on a number of resources like JavaScript files etc – at least the theme images are served directly from Apache so are fairly cacheable. Overall though the site is a pretty spectacular example of bad caching practices – just like it was when it was running on WordPress.

]]> By: Sam Ruby http://www.symphonious.net/2008/09/02/the-problem-with-atom/comment-page-1/#comment-173287 Sam Ruby Mon, 01 Sep 2008 21:09:11 +0000 http://www.symphonious.net/?p=949#comment-173287 There would not be anything invalid about producing an Atom feed where feed/updated is always time.now (as in, the time the feed was produced). A possibly related question: does IWWCM produce ETags and/or Last-Modifed dates, or does it recompute the entire feed on every request? There would not be anything invalid about producing an Atom feed where feed/updated is always time.now (as in, the time the feed was produced).

A possibly related question: does IWWCM produce ETags and/or Last-Modifed dates, or does it recompute the entire feed on every request?

]]>