A new major release of Movable Type - big things must lie ahead, given that so far, important new features like the plugin architecture came burried deep in the changelogs of minor releases. Reality might look bleak, though: ATOM API support, a CSS based backend and support for a new SixApart service called TypeKey - that's it.
TypeKey is a centralized single sign on system for comments on MT3 powered weblogs and thus, one of the last things one might need. Anil Dash: "We've built an entire service just so people developing cool things don't have to build login systems". Blogs don't need login systems. It is the same discussion we've had in the German blogosphere a while ago. OBKA, as it was called, is so dead that even the wiki where lots of people contributed their opinion is gone. If there is a need to control who published comments on my website, it is to fight comment spam. To do that, I don't need a login and registration system. I need a simple tool that gives me views like "all comments by that author", "all comments that contain this phrase or link" and "all comments from that IP address" on the mt_comment table. So if I as a user don't need a registration system, who does? SixApart does, of course.
Reason one might be TypePad. TypePad is not onyl a service that gets lots of things right. It also seems to have a core that scales much better than the MT core does - so that you can sell it, with custom user interfaces, to large customers like NTT. To those, a registration service might be a selling point. Reason two might be the user base such a system would generate when coupled to MT. It is unlikely that MT produces revenues. Commercial services that want to connect to the system do. Allowing such services to use the system does. If you called the same thing Passport, you'd be in trouble.
Does basically the mere support for TypeKey justify a major relase of Movable Type? No, if you connect version numbers with product improvements. Yes, if you connect them with a company's business strategy. Given the number and quality of improvements in earlier minor releases, a major upgrade with such thin news could be understood as a "Yes, MT stopped innovating" statement - but it will have a shiny new interface to distract from the fact that MT's basic weaknesses have not been addressed. With many active and technically advanced blogs existing long enough to have accumulated lots of data, these become more and more apparent. On top of the list:
- MT does not scale well
- The rebuild system is flawed
Both might not be a major problem if you run MT on a dedicated server. It is if you run it in a typical shared hosting environment where processing time is limited. If MT does not time out while rebuilding, it at least works generally slow enough to make the user interface for both commentators and backend operators frustatingly unresponsive. Addressing those issues might require to swap large portions of today's codebase, but could give MT the momentum back it has lost. Just patching in a service that some day might produce revenue by providing third party gateways won't do that.
Comments are closed at the moment. I will post a blog entry as soon as they are available again.
