Architecture and Design are terms often not heard enough within project teams – . It almost feels like these words where left behinds in the Computer Science universities. Two interesting books on the topic are:
What follows is a brief overview of the first book. Another blog posting will cover the 2nd book. Net out, I feel the Software Architecture in Practice book could have done more to embrace agile, and the concept of “done”
Book: Software Architecture in Practice
Page 4 start with a relevant definition:
Although a bit last month, Matt Turck offers an interesting read on the Bloomberg Terminal. A number of software packages are called out throughout the article, coupled with this possible key quote:
A data producer like [PsychSignal] publishes its social media sentiment index on [Quandl]. A user exports that data from Quandl to a high-end visualization service like [Plotly] or [DataHero]; extracts subtle insights using an analytics tool like [BigML]; backtests an investment strategy using [Quantopian], and then actually trades on that data using Quantopian again.
Interesting items mentioned:
Coupled of interesting posting on Docker usage
- Dockerfile Tutorial | Running Apache http server inside docker container
- Docker commandline Tutorial | Running Apache http server inside docker container
- Docker is the best fit for continuous delivery
“You can create container for each component of your technology stack and they can be used on developer machine as well as production environment.”
Interesting read over on Segments.io regards Server-Sent events. Main take away appears to be that if you don’t need full 2 way communications aka WebSockets, then Server-Sent events maybe the way to go. Anyone used Server-Sent events for a web trading application?
FT has an interesting article on Big Data. The closing paragraph summaries the article well:
“Big data” has arrived, but big insights have not. The challenge now is to solve new problems and gain new answers – without making the same old statistical mistakes on a grander scale than ever.
I’ve been waiting for a blog entry to appear for Mike Cohn’s recent monthly news article, “Paying the Cost for More Precise Estimates”. Unfortunately, apart from the campaign URL, nothing has appears on Mike’s blog as far as I can see. So without waiting any long, let me offer the summary view of an excellent article – estimation is not free! I realise this maybe obvious to some, but few manager understand the “cost” associated with estimation.
To add precision to an estimate requires time
So by far the coolest of these has to be the open sourcing of Apollo 11 Command Module code (Comanche054) and Lunar Module code (Luminary099)
More up to date, will WinJS make headway in the non Windows space?
To wrap up, there’s .NET Compiler Platform (“Roslyn“)
The .NET Compiler Platform (“Roslyn”) exposes a set of Compiler APIs and Workspaces APIs that provides rich information about your source code and that has full fidelity with the C# and Visual Basic languages. The transition to compilers as a platform dramatically lowers the barrier to entry for creating code focused tools and applications. It creates many opportunities for innovation in areas such as meta-programming, code generation and transformation, interactive use of the C# and VB languages, and embedding of C# and VB in domain specific languages.
Neeve Research offers an interesting product:
A platform to host, execute and manage business applications that are message driven, are collaborative and multi-agent in design and have extreme performance, availability and scalability requirements
VentureBeat offer an interesting on Agile in the Enterprise – agility more than agile, drop the agile fundamentalists. This leads nicely to the Bank Systems & Technology article, “Capital One Delivers 85% Of Software Through Agile”. Curious what version of “agile” Capital One is using? Innovation labs sound very sensible:
In order to encourage innovation through the company, Capital One rotates developers through the innovation labs, rather than just leaving team of developers in the labs full time, although Wolfs notes that there are a few associates who are permanently assigned to the labs. “My goal is make sure the labs are not the only place where innovation is done. Innovation should be encourages on all of the teams,”
Interesting read over on e-FOREX offers an interesting read on where MS Matrix is heading. Few screen shots in the article.
For example, a client may take advantage of our market leading pre- and post-trade Transaction Cost Analysis (TCA) on Matrix, but allocate through a third party. Matrix is our storefront, but behind it are robust services available to our clients’ particular workflow.
FDD has been around for some time. Although Scrum is more common, its not unusual to hear the Scrum theme described as feature. This leads to somewhat of a problem in some scenarios. Specifically, teams get driven down a feature road, where the Product Owner perceives all he wants is features, and begins to drive for features, sometimes at the mercy of accepting partially tested software within a fractured environment that bears no resemblance to the final deliverable state. Feature madness. Mentorship and education should be consider in these scenarios.
There has been a lot of chatter recently on why agile is not the word it used to be. I offer a few bullets on why agile has in some ways become an off-roading term for software development based on bad practices I’ve seen:
- Scrum daily stand-up’s are really either status meetings or team meetings. Nobody in the team appears to recall the three standard questions to be answered, or the time-box duration for the meeting
- 12 original XP practices are often lost by agile teams. How often have you seen an agile team start coding with no coding standards, no design, and no idea of the XP practices on testing
Interesting read over on Tradeweb’s blog on SEF, and its connectivity to MarkitSERV
If you take the SOLID principles to their extremes, you arrive at something that makes Functional Programming look quite attractive
SOLID, particularly the SRP and ISP, leads you towards code bases with many fine-grained classes with a single method. Such objects represent data with behaviour, but can also be modelled as behaviour with data: Closures. When that happens repeatedly, it’s time to make the switch to a Functional Programming Language like F#.
DirectX 12 look to offer reduced GPU overhead via features such as descriptor tables and concise pipeline state objects. New rendering pipeline features that will dramatically improve the efficiency of algorithms such as order-independent transparency, collision detection, and geometry culling.
Interesting read over on InfoQ, “The Case for a Dedicated Scrum Master”:
Software development teams generally catch on to the mechanics of the scrum process after 3 – 6 sprints, if they are properly trained and guided
Food for thought
From The New York Times:
Wall Street banks, facing tighter scrutiny from regulators, are moving to get out of the business of physical commodities trading.
On Wednesday, JPMorgan Chase took a big step in that direction, announcing that it had agreed to sell its physical commodities trading unit to the Mercuria Energy Group, a rapidly growing Swiss trading firm, for $3.5 billion in cash.
The Volcker Rule, part of the sweeping Dodd-Frank financial regulatory overhaul, restricts the ability of banks to trade for their own accounts. That extends to limits on some commodities trading.
LMEstage, a test environment that allows trading strategies to be tested against algorithms in a simulated live market.
Using custom algorithms that generate order flow, LMEstage enables clients to test their trading strategies and supporting systems in an environment that is similar to, but not identical to, the live market. In addition, the LME can stress the simulated market in a variety of ways, for example, by providing strong directional price pressure or high order rates.