It is possible that the select few will be motivated enough to free their own iPhone or create tools to detect violations of the end-to-end principle, but I worry that the critical mass will not be reached.
That's the beauty of capitalism – you don't need everyone to be looking for the best deal or even necessarily checking features or price in order for the system to work. If profit margins are relatively low, then you can't afford to lose even the minority of customers who actually look at the price/features of a product. (Supermarkets are good examples of this – I doubt that the majority of consumers check the prices on staple foods like rice, pasta, or salt.) Similarly, even though a majority of TV show-watchers probably watch their TV legally over the airwaves or cable, TV producers can't afford to lose the minority who are going to flock to alluc.org, and so they create alternatives – even though grandma probably wasn't going to watch alluc.org, she might watch hulu.com.
MikeRT
· 11 months ago
Open platforms are often better, but the input of many developers doesn't always have a good impact. Case in point, desktop Linux. Both GNOME and KDE are pretty bad as desktop environments and platforms compared to even Windows XP. GNOME is fairly light these days, but it's still a kludge to develop for compared to say Windows with .NET or MacOS X with Cocoa. KDE is such a bloated beast that I'm amazed that anyone still uses it.
I put the blame for these things on the fact that you don't have enough central planning. Contrary to popular opinion, complex systems do need to be designed, not thrown together by committee. I'm not saying that you are advocating that, but a lot of people underestimate the advantage that the less open platform vendors have with being able to have a systematic design for their platform. That tends to result in a great deal of consistency.
Java's method of community involvement has worked out quite well. That sort of thing should be the future of how open platforms are worked on. There does need to be some serious leadership to ensure that the platform is well designed and not allowed to morph into a frankenstein.
Tim Lee
· 11 months ago
Yeah, i definitely agree that there can be too little central planning as well as too much. However, I do think that once an open platform becomes dominant in a particular market, it's extremely unlikely to get displaced by a closed platform.
MikeRT
· 11 months ago
That's very true. One of the best examples that comes to mind is blogging software. WordPress has destroyed the market for closed solutions here, which is part of the reason why Movable Type now has an open source distribution. It's a sad state of affairs, actually, as a quick look through the capabilities shows that WordPress is actually a very, very inferior platform from a developer's perspective. It's like comparing Linux to MacOS X.
Fortunately, working on plugins and such for blog software is just a hobby of mine, but it would be basically impossible for me to make a living writing plugins for either platform because frankly, no one wants to actually buy anything like that anymore. They've grown accustomed to people giving away their work.
Overall, I think open source isn't bad and am firmly in agreement that open platforms are the best way to go, but my experience in the DC area has taught me a lot about the dangers of having an engineering field that relies very heavily on support and contract services. I would hate to see a future where consulting is the main way that people make money off of software development because it would be a future of low investment into R&D (IT contractor corporations make far less profit than product companies of the same size).
Ryan Radia
· 11 months ago
Out of curiosity, what's inferior about WordPress and what do you recommend instead?
MikeRT
· 11 months ago
I was a bit harsh on WordPress. It's not a piece of crap, but it's also not exactly a shining example of good software engineering either. My beef with it is that it's like a lot of PHP applications... it's a mess of PHP, HTML and SQL. The reason I prefer Movable Type is that it's got a very clear separation in that respect. For example:
1) All basic data access is through an object-relational mapper similar to Hibernate for Java. I've never had to write a single line of SQL to do things like load entries or comments from my database, and defining new tables and behaviors is as simple as defining a new object.
2) Movable Type has a much more sophisticated template system based on XML that fits in naturally with HTML.
3) Movable Type has a lot more developer features, and they're significantly better documented than WordPress is if you go based on official documentation.
4) Movable Type's CMS is actually an application built on the MT framework and template system; they eat their own dog food as developers which shows how powerful their framework really is.
5) It's just a lot easier to extend for cool stuff. All of the hooks into it are straight forward, and adding new features is generally very intuitive from a developer's perspective. I've wrote a plugin for WordPress once, and found it to be a lot more painful to get the basics going than was the case with Movable Type.
6) Perl is a more developer friendly language than PHP once you get your mind wrapped around basic Perl. Ideally they would have used Python, but Python is too often barely supported on shared hosts.
I respect the work of guys like Matt Mullenweg, but at the same time, there is a real cult around WordPress. I mean seriously... it can't even handle two blogs from the same installation. They had to create a fork of it to handle that.
All Comments
That's the beauty of capitalism – you don't need everyone to be looking for the best deal or even necessarily checking features or price in order for the system to work. If profit margins are relatively low, then you can't afford to lose even the minority of customers who actually look at the price/features of a product. (Supermarkets are good examples of this – I doubt that the majority of consumers check the prices on staple foods like rice, pasta, or salt.) Similarly, even though a majority of TV show-watchers probably watch their TV legally over the airwaves or cable, TV producers can't afford to lose the minority who are going to flock to alluc.org, and so they create alternatives – even though grandma probably wasn't going to watch alluc.org, she might watch hulu.com.
I put the blame for these things on the fact that you don't have enough central planning. Contrary to popular opinion, complex systems do need to be designed, not thrown together by committee. I'm not saying that you are advocating that, but a lot of people underestimate the advantage that the less open platform vendors have with being able to have a systematic design for their platform. That tends to result in a great deal of consistency.
Java's method of community involvement has worked out quite well. That sort of thing should be the future of how open platforms are worked on. There does need to be some serious leadership to ensure that the platform is well designed and not allowed to morph into a frankenstein.
Fortunately, working on plugins and such for blog software is just a hobby of mine, but it would be basically impossible for me to make a living writing plugins for either platform because frankly, no one wants to actually buy anything like that anymore. They've grown accustomed to people giving away their work.
Overall, I think open source isn't bad and am firmly in agreement that open platforms are the best way to go, but my experience in the DC area has taught me a lot about the dangers of having an engineering field that relies very heavily on support and contract services. I would hate to see a future where consulting is the main way that people make money off of software development because it would be a future of low investment into R&D (IT contractor corporations make far less profit than product companies of the same size).
1) All basic data access is through an object-relational mapper similar to Hibernate for Java. I've never had to write a single line of SQL to do things like load entries or comments from my database, and defining new tables and behaviors is as simple as defining a new object.
2) Movable Type has a much more sophisticated template system based on XML that fits in naturally with HTML.
3) Movable Type has a lot more developer features, and they're significantly better documented than WordPress is if you go based on official documentation.
4) Movable Type's CMS is actually an application built on the MT framework and template system; they eat their own dog food as developers which shows how powerful their framework really is.
5) It's just a lot easier to extend for cool stuff. All of the hooks into it are straight forward, and adding new features is generally very intuitive from a developer's perspective. I've wrote a plugin for WordPress once, and found it to be a lot more painful to get the basics going than was the case with Movable Type.
6) Perl is a more developer friendly language than PHP once you get your mind wrapped around basic Perl. Ideally they would have used Python, but Python is too often barely supported on shared hosts.
I respect the work of guys like Matt Mullenweg, but at the same time, there is a real cult around WordPress. I mean seriously... it can't even handle two blogs from the same installation. They had to create a fork of it to handle that.