Skip to main content

Programmers who "care" vs $2-shop software.


There is something to be said for the experienced coder who is capable of producing low quality software at a fantastic rate. As a programmer who cares about software, it can be easy to get caught up in the idea that software must always be thoroughly tested and beautifully architected, appropriately apply design patterns, separation of concerns, and all the rest of it.

Whereas sometimes, a business needs a rough-and-ready, potentially high maintenance solution, until such time as either a venture stabilizes or, alternatively, crumbles away.

[EDIT 20121030: in New Zealand, we have a chain of shops called "The $2 Shop" (http://www.2dollarshop.co.nz/), which sell very cheap stuff that doesn't last very long but is good if you're in a hurry]

The work of a $2-shop programmer is often ruthlessly ridiculed by programmers who "care" (that is, programmers who are accustomed to applying a quality/craftsmanship oriented approach to their work). The $2-shop programmer however relentlessly and undeniably produces lots of working software - it's just that it's difficult to maintain, and augment.

$2-shop software however - whether it's written by a keen business owner, or by an opportunist contractor who learned VB from a web tutorial - can often end up being the core of a bigger system. This is because, as a business venture stabilizes, it attracts higher quality (more stable) programming, resulting from the desire to make the process more robust ans scalable, for the longer term. The trouble is that businesses may not be inclined to significantly refactor existing software infrastructure and instead opt for high quality clip-ons to an existing, $2-shop core.




Herein lies the problem. If a business wishes to attract quality programmers, it needs to be open to the idea that they are most likely going to want to refactor $2-shop code when they come across it, in the same way that if you were investing in extending a house you would want to address hazardous electrical wiring in the process.

If a business with pre-existing $2-shop software infrastructure (let's face it, this is most of them) is willing to take this approach - that is be open to investment in significant refactoring, and trust it's programmers to advise as to where this is appropriate - then it (a) introduces a short-term cost and (b) sets itself up for long term stability.

In summary, I think that there is definitely a place for $2-shop programming - it is in the realm of prototypes, high risk ventures, and where there are significant gains to be made by acting very fast. And for the resulting $2-shop software to become part of a sustainable and scalable business model, businesses must accept the cost of refactoring.



Comments

Popular posts from this blog

HOW-TO: Add/edit a field in Team Foundation Server 2012 using Visual Studio 2012

It's been a while since I made a purely technical post...

So, today I wanted to make a change to a Microsoft Team Foundation Server 2012 (TFS2012) instance that I am working with to reflect "Actual" time spent on a task - mainly for reporting purposes, and because I have found in the past that making this minor process adjustment yields a relatively useful metric over the long-term.

I am using the Microsoft Scrum 2.1 Process Template (http://msdn.microsoft.com/en-us/library/vstudio/ff731587.aspx) for a project that I am working with. So that I don't forget how to do this (again!) I will blog-post the procedure I've used to add this field to the template as a screen-shot-based tutorial, as follows...
Before Assuming you are familiar with the Scrum Process Template (2.1-ish) - open a task and take a look at the "Details" section, as follows:


 This is where I want my "Actual" field to show up.
Get the Power Tools Download and install the latest v…

HOW-TO: Apply a “baseless merge” in Team Foundation Server 2010 (and 2012)

Another purely technical post on TFS...
The scenario We wish to migrate code between branches that do not have a branch/merge relationship, in order to expedite urgent changes being made by a project team, without disrupting on-going BAU development work. Sample branch hierachy/strategy Imagine the following branching strategy in TFS (visible by connecting to TFS via Visual Studio 2010 or 2012):

Essentially you have a "DEV" branch, which has a "QA" branch, which in turn has a "PROD" branch. DEV is the branch that you would be using for BAU development. As a piece of development matures, you move it into QA, where it is tested by your internal QA team. There may be further changes made in DEV that are moved into the QA branch as the QA team pick up issues. Once the QA team are happy with a packaged of changes, they will move them into PROD, which is essentially the hand-over to the customer. The PROD branch represents the software that the customer has.

Mo…

Eclipse/Android error: "Multiple dex files define [...]"

Wow, I am really going nuts blogging this-evening - 2nd post in less than an hour. 


Anyway this is a particularly nasty error that I keep running into with Eclipse/Android when starting the emulator after I have not run it for a little while. Since I run the risk of permanently forgetting the solution to the problem every time I walk away from my Android project (and thus having to spend a painful hour-or-so digging up the procedure again), I will blog it here, for my benefit, and for the benefit of anyone who may also suffer the same problem.


The gist is that when you start the emulator in debug mode (that is, you hit the button in the following image), you get the following error message come out on the console and a nasty popup telling you nothing more than there is an error with your program and you need to fix it:








[2012-04-06 23:20:57 - Dex Loader] Unable to execute dex: Multiple dex files define Lcom/google/gson/ExclusionStrategy;
[2012-04-06 23:20:57 - SimpleList] Conversion to Dal…