Apolon's Software Torq

I decided to add my name to the title of the blog. This somewhat selve serving, but hey it makes it easier for people to “put a name to the face” where the blog is my “face” on the web. I read a bunch of blogs and when the blog title is too obscure or the author’s name doesn’t come through anywhere on the blog reader, there is no association with the author in my mind. So… when I come across that name somewhere on the web, there isn’t an automatic link in my mind to their blog. The mind link to a blog provides heaps of context for who they are, their default position on a bunch of subjects etc.

Posted by Apolon at 11:40 AM | Permalink | Comments | links to this post

My interpretation of Microsoft’s vision for Sparkle is as a tool for designers while developers will use Visual Studio and they both will work on a Visual Studio project/solution. This is a focused approach but I’d love to see Sparkle spread it wings somewhat and be used in the varied scenarios that people dream up when given a tool with extensibility. The extensibility points that would be beneficial include:

• An extensibility API.
• Sparkle addin support
• The ability to extend the user interface with scenario specific functionality.
• The ability to define your own data binding source and a UI that supports that data source.
• Not having to be tied to a Visual Studio project to edit XAML files.
• The ability to embed scripts in XAML files instead of compiled C#/VB. Maybe this relates to the WPF/E plans, but allowing an arbitrary scripting language instead of JScript specifically would be beneficial.

Unfortunately, I don’t have the time currently to experiment with Sparkle in depth. Paying work comes first, but man it looks promising.

Posted by Apolon at 11:10 AM | Permalink | Comments | links to this post

I’ve just downloaded and ran the Sparkle January CTP. Congratulations to the Interactive Designer team for their first public release!

One of the first things that was noticed was the start up time took ages – well at first. Subsequent runs came up quickly (around 2 seconds on an Athlon XP 3200). To reproduce the slow startup, the machine had to be rebooted and Sparkle took around 23 seconds to start. There’s obviously a bunch of DLL caching going on as per usual with Windows – but the difference in startup time seems even more pronounced that a typical Windows app. The install directory for Sparkle January CTP shows up a ton of really tiny DLLs, which is surprising since I thought that the current wisdom is that fewer larger DLLs is better for startup performance. This is probably just a temporary scenario for the CTPs.

It would be great if the Sparkle development team could comment on WPF application design from a performance perspective and some of the design/deployment decisions and tradeoffs made while developing a decent size WPF application such as Sparkle. One of the issues moving to a newer software technology historically is that it always seems to need bigger processors, more memory etc. This makes it more difficult to get users to upgrade regardless of the benefits to developers (self centered as we are). Getting some .NET/WinFX performance related feedback relating to the first publicly available “largish” WPF application that I know of would be invaluable.

PS: Yes I’m calling it Sparkle. It just takes too long to type Microsoft Expression Interactive Designer!

Posted by Apolon at 10:23 AM | Permalink | Comments | links to this post

Here’s a line from Paul Allen’s Official Provo Labs Announcement post that came through in my river of news last month :

“In the coming months we’ll be looking for part and full time developers, internet marketers, content experts, sales people and interns and lots and lots of contractors.”

The last part of the sentence immediately conjured up Scott Sigler’s voice in my head doing one of his signature podcast novel intros:

Welcome to the bubble (1.0, 2.0, …) project <insert-your-project-name-here>, … This project contains mature situations, language and lots and lots of violence….er I mean contractors.

Maybe its just bad experience from the dot com bust, but the term “lots and lots of contractors” sends shivers down my spine. I wasn’t in a position where the costs of paying for “lots and lots of contractors” affected me. In fact I was a contractor, but did see how quickly such a situation can drain the life blood from a company. Software/tech contractors are typically highly skilled and capable people, but it would be wise to question the benefit of having a large part of your workforce made up of individuals who have only a short term interest in the success of your company or project.

Posted by Apolon at 9:24 AM | Permalink | Comments | links to this post

I’ve been aware of the "Getting things done" approach to time management from mentions on various blogs and podcasts. Today I came across the reference again and decided to maybe buy a book on the subject. Amazon listed hardback, paperback, audio cd, audio casette and and audio download. The audio download was at a reasonable price of $9.95. Now I’m of the opinion that e-books and audio downloads should be significantly cheaper than their physical counterparts such as books and CDs. Its frustrating when you see a price only marginally cheaper for the electronic version of some content when the seller doesn’t have to allow for the physical manufacturing costs, the cost of stocking the physical items and the cost of moving those items around.

When going ahead with the audio download purchase, I got shuffled over to audible.com. This required me to log into audible and then find the audio download again. An attempt to buy the audio came up with this:

Huh? Don’t you want my business? Now this may be due to restrictions placed on audible.com by the author, but from a customer point view, I just wanted to act on an impulse purchase without much hassle. Now I can’t really be bothered.

Posted by Apolon at 4:14 PM | Permalink | Comments | links to this post