Our weekly reading found a lot of interesting posts, this past week, so making a choice on the posts to include was quite hard. Again, some posts were not originally posted this week, but they are interesting enough. So let’s begin.

• Richard Morris’ Developing for Delivery, a Practical Example addresses the difficulties of keeping a database current, when the database exists in a multiple number of sites, in a wide range of versions and how such difficulties were addressed by Calvi, a provider of telecom invoicing software. From personal experience, this is not an easy thing to achieve, even at a smaller scale, and the article provides interesting advice on how changes in processes and the use of adequate software tools can help reduce the difficulties involved.

• Alex Bolenok’s NULL in SQL: explaining its behavior is a very good article on the idiosyncrasies of NULL behavior in SQL. The use of NULLs is not at all clear, especially for developers less familiar with the workings of databases, so the article is almost of mandatory reading. Most interesting.

• Microsoft’s SQL Server Development Customer Advisory Team’s SQLCAT’s Guide to: Relational Engine is a free ebook that includes relevant posts from SQLCAT’s blog, from 2005 to 2012. Recommended.

• Rob Farley’s Spooling in SQL execution plans is a few month’s old post, but one that clearly shows why spools are used in SQL Server execution plans and why they are not that bad.

• Alex Popescu’s The premature return to SQL is a response to one article include in last week’s links, by Jack Clark, in which the author explains why the premature return to “SQL” is wrong – this “premature” return is motivated by an attempt to capture financial gains, does not consider the fact that many NoSQL products, in spite of having not yet reached technical maturity, have already provided valuable, alternative new doors to data and results, basically, from pressure from database vendors. 

• Uncle Bob Martin’s Dance you Imp’s is a humorous but no less interesting article on Object Relation Mappers and the impedance mismatch between OO and the RDBMS storage used to persist them. In a very funny way, the author concludes that there is actually no object to relational mapping. A very interesting read.

• Jimmy Bogard’s Scaling lessons learned–from 0 to 15 million users describes the lessons learned while building a system that has grown to handle up to 15 million users over the last 3 years. It is an excellent read and the lessons can be of use to any who develops systems that need to be able to scale (and even for those that don’t have such a need).

• Sean Hull’s 20 Obstacles to Scalability addresses some key points to consider, when designing a web application that will need to scale. Although based on a MySQL based web app, the advice is general enough and applicable to any RDBMS based web application.

• James Turner’s What Developers Can Learn from healthcare.gov addresses some of the issues exhibited by the healthcare.gov website, to draw more general lessons regarding load testing, good looks vs. functionality, and validation, that can be of value to any website. A very good read.

• Chris Andrè Dale’s Why it’s easy being a hacker – A SQL injection case study, although from last January, addresses the issue of SQL injection vulnerabilities and the fact that many easily available teaching materials used by developers may actually contribute to the persistence of such vulnerabilities. It is a very good read and it draws the attention to one issue that can, indeed, have negative consequences – the  influence of teaching materials on developers and their work. This is an issue that will deserve a future post here.

• The last link for this week is Peter Krantz’s Quotes from the NATO Software Engineering Conference in 1968. It’s a link from 2011, but it’s so very interesting because it shows how things don’t really seem to change, even after so many years.

 That’s it for this week. Thanks for reading.