techno.blog("Dion")

NOTE: It goes without saying, but to be clear, none of my thoughts here are related to Mozilla itself

Mitchell Baker has a post on the EC prelim conclusion on Microsoft and IE:

From EC: “Microsoft’s tying of Internet Explorer to the Windows operating system harms competition between web browsers, undermines product innovation and ultimately reduces consumer choice.”

Mitchell has some very interesting points, and the two high level ones that stick out at me are:

They did a bad bad thing

In my mind, there is absolutely no doubt that the statement above is correct. Not the single smallest iota of doubt. I’ve been involved in building and shipping web browsers continuously since before Microsoft started developing IE, and the damage Microsoft has done to competition, innovation, and the pace of the web development itself is both glaring and ongoing.

Bloody hell, coming up with a remedy that is fair and good for the Web is tough!

There are separate questions of whether there is a good remedy, and what that remedy might be.

…

The extent of the damage is so great that it makes it difficult to figure out an effective and timely remedy. I believe it’s worth some effort to try.

Thinking about the users, you end up at PPK’s point: 0 or 5. Hard indeed.

Also, take a good look at the entire Mitchell post as it is really interesting to see how Firefox has grown despite of the barriers, and that Microsoft shouldn’t be able to use it as a “see, look what can be done.”

Bizarre times indeed.

I have had the pleasure to be at PDC this week and Microsoft put on a great show. As they showed their vision of unification around Windows (cloud, Web, PC, mobile) through great developer tools, there was excitement.

Windows Azure looks great. The “on premise” feature looks particularly intriguing. If they can bridge the data center and the cloud, they have something quite compelling. Enterprises are struggling with the cloud in part. What do you put up there? How do you secure it? How do you tie back? Microsoft is going after that problem.

I am curious about the details. Ray talked about how you get to “leverage your skill set”, and showed mappings between the Windows platform and the Azure one (SQL Server – SQL Services). How similar are they? Surely there are limitations when you are in the cloud (App Engine has restrictions on request timeouts, file sizes, threading, and other APIs). Running unmanaged code? Virtualization is good enough that you can just do that these days huh. Impressive. Again, what are the limitations? You can imagine people deploying platforms if this is Open enough. E.g. platforms that are currently in C, or C#, Ruby (IronRuby), Python, (IronPython), …

Have they got SQL Server so it “just scales”? I noticed that the demos that ChrisAnn and Don Box did with SQL services dealt with XML tuples of data when inserting data, and LINQ to get data out. I am excited to hear more of the details.

In fact, I feel that way about a lot of the things that were released. I want to dive in more. Many of the talks stayed very high level indeed. Ray Ozzie was a good speaker, but spent a long time talking about the history of computing to put things in context. I get it, they have to do that for the press etc. They are setting a tone of “this is where we are coming from, this is where we are going.” For developers though, I wish for once someone here, or at JavaOne, or at MAX, (whatever) would stand up and say “Ok, let’s build some amazing things in front of you and show you what we have done with our tools and platforms.”

Don and Chris do that in their talks and people love it. They were on form as always this time around, although I wish that instead of showing AtomPub, they showed us Oslo, M, etc. I know that Douglas talked on that later, but I would have loved to have had that content in the Don/Chris show. I enjoyed it, but when it was done I thought “wait, did I just sit here for an hour watching manual HTTP?” Don’t get me wrong, showing the path from local development, to running on your on premise server, to the cloud is compelling, but I wanted more :)

Game On

Everything that was announced we pretty much knew was happening. From the cloud, to the Web (Silverlight 2 / IE 8), and beyond. But, even though we knew about this, I don’t know if we thought they were this far along. Microsoft is executing. This show set the stage “this is where we are going, and look how far we have come.”

The Office on the Web demo showed that. Works in all browsers, with enhanced Silverlight support. Very nice indeed. What a wake up call to the rest of the Web?

And, what about tools. The WPF version of Visual Studio 10 looked fantastic. Having a plugin model that makes it so easy to interact with the editor looked snazzy and useful indeed. I have long wanted to separate the source code from the visualization of that same code. The source code in the repository can have curly braces on separate lines, but I will see function() {. In my world, comments in the style of “<author>Dion Almaer</author>” will get converted to “Author: Dion Almaer” and made small.

Blue Pill

If you are a Microsoft developer you are probably very happy at Ray Ozzie’s coming out party. Others on the fence may be interested. For those of us who worry about handing Microsoft control of the browser, plugins to other browsers, the cloud, the server model, and more…. I won’t lie to you. I am cautiously observing. Silverlight adoption worries me.

We can’t fight Microsoft with “don’t choose them, remember what they did to you before?” Fear is lame. Instead, this is a wake up call to Adobe, Google, Yahoo!, Amazon, IBM, Sun, [insert other developer / platform players] to get kicking.

We can’t just be Open, we have to be better!

Bruce Eckel pointed to Microsoft Spectra, the new feed reader that uses Flex, and he questions their lack of dogfood eating. Of course, this is MSNBC (which is far from Microsoft), and who knows when development started.

What I find more interesting is the fact that beyond the flash-y visualizations, would you actually want to read news like this:

Rich visualizations are great when they show you something in a new way that adds value. This just seems to make it harder to get to the content :)

Even Ryan Stewart must think that this isn’t a good “RIA” ;)

Microsoft stopped IE and messenger. Roz got on stage for years promising a native Office, and then when it gets here it is disappointing.

I like some of the MS stack. Expression Studio, Visual Studio, Silverlight, .NET, all promising.

It could have been for good reason (e.g. Safari really was decent then…) but, what about the trust?

And, when I see people the Olympics and the Library of Congress even signing up for Silverlight content? Erg.

It is fantastic to see IE8 up the ante on concurrent connections from 2 to 6. I would love to know how they made the call for 6. It sounds reasonable. You don’t want too many else you could end up sucking down so much content at once it could freak out the browser (e.g. imagine a bunch of video on a high speed line).

This should mean, for some situations, that you may not need to deal with www1, www2 type hacks to split up your domains. Of course, sometimes you will still run into the limit and it will be required.

I do wish that Twitter would up their ante too. Maybe not by changing the 140 char limit itself, but maybe by supporting #hashtags so they don’t “take up the room” and instead get moved down below to the metadata layer etc.

The IE 8 updates are out, and there are some great features that relate to Gears. Also, a lot of the features are standards based so we will see the functionality in other browsers too.

Six connections per host

Yay! The 2 connection limit has been so painful for rich Ajax development, especially when you get into advanced work like Comet. This small change is going to be huge.

Connectivity Events

Connectivity events allow websites to check when the user is connected to the network and receive notification of connectivity changes.

This is fantastic. A lot of developers want this functionality, and we have held off implementing it in Gears as it is actually quite tricky to do correctly. Having a version of this in the browser will be great, and Gears applications can leverage it.

DOM Storage

The simple storage for offline applications. I could see someone writing an app that uses DOM storage for simple cases, and Gears for the more advanced (SQL vs. name/value pairs).

Cross Domain

Cross domain is actually going to happen in 2008, which is very exciting indeed. By jumping in we can iron out the issues and end up with a real mashup world that is more than read only tools. I am assuming that this is also using standards such as postMessage.

Cross-domain communication is an integral part of AJAX development
and mashup Web applications. Internet Explorer 8 includes two
features to help you build cross-domain communications that are safe
and easy to implement:

• With Cross-domain Request (XDR), developers can
  create cross-site data aggregation scenarios. Similar to the
  XMLHttpRequest object but with a simpler programming model, this
  request, called XDomainRequest, is the easiest way to make anonymous
  requests to third-party sites that support XDR and opt in to making
  their data available across domains. Three lines of code will have
  you making basic cross-site requests. This will ensure data
  aggregation for public sites (such as blogs) will be simple, secure
  and fast.
• Cross-document Messaging (XDM) APIs allow
  communication between documents from different domains through
  IFrames in a way that is easy, secure and standardized.

I was gutted to read that Dare is quitting his blog. He almost did it before.

This would be a real loss, as Dare is someone inside Microsoft who gives his honest voice and opinion to the community. I am sure that people within the company aren’t always pleased at what he says, but he is doing everyone a great service (including Microsoft).

I have a couple of hopes:

• That Dare comes back with renewed vigour, refreshed, and ready for more
• That he isn’t quitting the blog due to someone being a pain in the arse and getting him to quit.

At first we had Nelly and XSL/T, and then we had Lisp. Now it is time to to evolve to C^HJavaScript.

ps. There is nothing geekier than a ^H joke.

At first we had Nelly and XSL/T, and then we end up with Lisp.

Paul Krill asked Bill Gates about declarative languages and how they are the future (and ironically the past!). Bill talked about some of the work happening at Microsoft:

“Most code that’s written today is procedural code. And there’s been this holy grail of development forever, which is that you shouldn’t have to write so much [procedural] code,” Gates said. “We’re investing very heavily to say that customization of applications, the dream, the quest, we call it, should take a tenth as much code as it takes today.”

“You should be able to do things on a declarative basis,” Gates continued. But this has not caught on partially because of weak data models — first Codasyl and then relational. Stronger data models since have emerged, such as rich schemas around XML as well as modeling work being done by Microsoft and others, Gates said. “We’re bringing the data models up to be much, much richer, and we think in that environment, a lot of business logic can be done in a declarative form. Now, we haven’t totally proven this yet. We’re doing a lot of internal developments ourselves that way,” including some Microsoft business applications, he said.

“We’re not here yet saying that [a declarative language has] happened and you should write a ton less procedural code, but that’s the direction the industry is going,” Gates said. “And, despite the fact that it’s taken longer than people expected, we really believe in it. It’s something that will change software development but more like in a five- to eight-year timeframe than overnight,” he said.

I am guessing that the work of Doug Purdy, ChrisAn, and Don Box is fitting into this world, as well the mentioned work of Brad Lovering.

But…. maybe the future is a port of Jelly ;)