Knowing .NET

Intel’s announced the availability of a prototype C++ compiler implementing Software Transactional Memory, one of the central topics of this brilliant .NET Rocks show.

That’s exciting, although I suspect that on just two cores STM will be disappointing.

If you’re eager to get your first glimpse at Visual Studio X (you heard it here first!), you can download a VPC image Visual Studio

Huh. I just received a link in email to “my” August issue of Dr. Dobb’s Journal. I’m not going to post the link I got, since it’s undoubtedly linked to me, but can anyone access http://www.nxtbook.com/nxtbooks/cmp/ddj0807/index.php?

It’s a Flash-based interface, but even when viewed in profile, full-screen (1050 x 1680), it’s annoyingly fuzzy:

Which zooms in as:

There doesn’t appear to be any way to control the antialiasing within the Flash interface. Blech.

Come and get it

P.S. I really, really like this use of VPCs.

John Lam details the first public availability of IronRuby. Couple reasons why I’m interested in this:

• It’s Ruby
• It’s the CLR
• It’s a second data point for how to code for the DLR

I don’t think I’m going to be able to resist the temptation to write a compiler for the DLR. I know I should resist, but I spend so many cycles thinking about programming languages and the DLR seems to have so much promise to language implementors.

Argh, I can’t believe I have such a busy week in front of me.

Wesner Moise’s NStatic static-analysis tool for .NET appears to be approaching its initial beta. Moise has made a number of exciting claims for this technology since he began discussing it about 18 months ago. If I understand correctly, NStatic involves considerably “deeper” analysis than most quality-assurance tools; it almost seems it applies the field of constraint-based programming to parsed program structures. At least, that’s the only way I can get my head around this post.

You know I love my homeys at SD Times, but ought not the headline read: “Study states obvious, costs money”?

I’ve fiddled around with the examples of GLEE, Microsoft’s layout engine for graphs and wish I’d had it for some visualization projects I did a few years ago. Looks like a nice library to add to your collection.

Alan Zeichick, commenting on the release of Oracle DB 11G Beta ranks the major databases as:

1. Oracle
2. DB2
3. SQL Server
4. MySQL

I’ve never been a fan of DB2, but Oracle DB is a rock and I would never hesitate to recommend it. (This despite my involvement with Oracle’s abysmal XML-based file-system (FS? FX?) ). Of course, many .NET shops use SQL Server and I’ve yet to experience a real problem with that. MySQL is preferred for Rails development and, again, I’ve never had a problem using it.

Based on comments to my earlier post, I put further work into Crystal Reports XI and SQL Server Reporting Services. Both seem capable-enough in terms of handling the data volumes and seem to have similar ease-of-use as far as the ultimate end-user goes (the inexperienced programmers who will be charged with writing reports). My take is that Crystal has more flexibility in terms of output formats, which might influence my choice.

However, there’s another wrinkle. The data is stored as an XML blob in the database. These are fairly hefty documents, 1-4MB apiece, with quite complex schema. But I don’t want to report on a single one of these blobs, I want to report on all of them. So my data model is not-quite-XML and not-quite-relational.

For instance, to use the cliched example, every record in the database contains data such as:

<?xml version='1.0' encoding='utf-8'?> <book> <authors> <author>Arnold Aarnison</author> <author>Bob Billson</author> </authors> <sales>23</sales> </book>

And the reports we’re looking for might be “Total number of sales of books by author” which would seem to necessitate either:

1. Joining the data in multiple database rows (i.e., make a new XML document that wraps the per-database-row <book> element in a root element called, say, <books>); or
2. Mapping the XML schema into a relational view (i.e., create tables for Books, Authors, and Sales); or
3. Reporting that can somehow combine SQL and XPath in a single query (e.g., “Select count(myXmlColumn.Xpath(/book/sales/text()) from myTable group by myXmlColumn.Xpath(/book/authors/author/text()”)

Option (1) is easy enough, but the resulting document is going to be massive — gigabytes. I’ve never worked with XML documents of that size and worry (prematurely?) about the capacity / performance of feeding a report generator such a volume of XML.

Option (2), which is essentially the reverse of (1), would solve the reporting problem (since once it’s in a relational structure, any reporting tool will work) and is do-able using Altova’s XMLSpy “Create DB Structure from XML Schema” but I’m concerned about the amount of manual labor required (I have to define dozens of foreign keys relationships by hand — but is that a one-time cost that’s no big deal?).

Option (3) would be the ideal, but I don’t think there are any reporting tools that work that way.

Any real-world experiences with similar challenges would be appreciated…