« July 2005 | Main | September 2005 »
August 31, 2005
I think there's a modern art exhibition in here somewhere...
I wasn't going to post another of these (seemingly endless line of) Bizarro-world RAD dialogs, but was spurred on by this and didn't want anyone to think that RAD was resting on it's laurels.
(Thanks to Marty to supplying this)
Posted by Andy Marks at 05:54 PM | Comments (0)
August 21, 2005
"... if anyone orders Merlot, I'm leaving!"
Many people would recognize that quote from the film Sideways and whilst I omitted the more colourful and memorable statement that follows immediately, I think this one captures the moment quite nicely. And in the same vein, I'd like to pose this hypothetical:
Assume you're applying for a development position at a new company: the interviews have gone well, good vibes on both sides, no red flags as yet, conditions good, people nice, pay acceptable, full steam ahead... and then you strike the iceberg when your potential employer (apropos of nothing) mentions that they are using software product X for whatever product X is intended...
Your mouth drops open, your pupils dilate and your struggle to hold back the rising bile caused by hitherto carefully buried memories of past encounters with the self same product X. Mustering as much decorum as possible given the gravity of the situation, you rise from your seat, carefully button your suite jacket (remember: "sometimes, always, never"), spit in the general direction of your almost-but-not-quite-employers and leave the room at a rapid trot.
Could you see this happening to you - with or without the flem? Have you encountered any products that inspire a revulsion that would lead to such a deal-breaking situation? I know of many people who have voiced such sentiment about a certain popular email/groupware client and whilst I share their pain up to a point, it wouldn't have that sort of effect should I need to use it again. However, I think I may have come across a legitimate contender in a version control system that is so unsuited to, well, version control that it's ongoing existence beggars belief.
This realization was a shock to me as I place product choice a distant last on the list of determinents of a successful development project - somewhere behind people, people, people, management and customer support, process and planetary alignment. So I was somewhat surprised to find myself reacting so strongly to this tool and it's constant ability to derail my struggle towards some semblance of decent productivity.
But enough of my gripes - I'm keen to hear whether other people have a list (whatever size) of prohibited products that would stop them knowingly working in any environment inhabited by said products.
As a further constraint to my hypothetical job position, assume there is little scope to change/replace product X, irrespective of your ability to mount a credible argument for it's termination based on technical merit alone. As I've voiced recently, many of these choices are made for reasons entirely outside the technical realm and the worker bees have no choice but to "like it, or lump it".
Posted by Andy Marks at 11:44 AM | Comments (8)
August 16, 2005
There's nothing like a nice informative dialog... Part II
Allow me to present another in this season's line of RAD dialogs.
Posted by Andy Marks at 12:07 PM | Comments (0)
August 11, 2005
There's nothing like a nice informative dialog...
And this is nothing like a nice informative dialog!
(Image reprinted without permission from (Ir)Rational Application Developer.)
Posted by Andy Marks at 01:42 PM | Comments (1)
August 07, 2005
A Case Of "Too Little, Too Late"?
Riddle me this, Batman: When is a software tool that costs literally nothing to obtain a poor purchase?
Answer: When the cost (measured in either time or sheckels) to install, configure, maintain, understand, use and or work around far exceeds the cost of using other solutions... Yes, this is an example of those tedious Total Cost of Ownership (TCO) theories that get so much airtime in much publicised Battle of the (Vendor Sponsored) Analyst Reports about whether, for example, Windows or Linux is a better choice for your common-or-garden-variety corporate IT shop.
However, this time I'm attempting to relate this to one of my pet gripes about software consultancy: we often seem to be engaged too late in the project to make the level of changes we are capable of. Specifically, I'm referring to a shortsighted and/or ignorant choice of development tools made well in advance of the formation of the development team, often with little/no practical knowledge of how these tools will impact the development process.
Before I continue, I should state for the record that I have no desire to mention any products by name. Nor is this a sideways swipe at OSS, although the "costs nothing" statement might have you thinking in that direction. Truth be known, the inspiration for this article comes from exposure to several high profile (and typically high cost) products, the combination of which has placed more obstacles to productive software development than any number of Mythical Man Month scenarios you might care to document.
The "choice" of these tools came about via corporate decree (i.e., existing licence agreements) and the "would you like fries with that?" approach of large product companies throwing in free copies of ToolMaster 3000 with every kids' meal. The practical upshot has been a development effort that is producing code at least 50% slower than it should, IMHO.
And in a lot of environments, these tool choices (e.g., version control system, IDE, application server) are effectively defaulted from what has been used in the past, with the development team having to accept the incumbent toolset with little option but to make do the best they can. And although tool choice is almost never (solely) responsible for a projects' demise, it can certainly make a large impact on the speed with which a development team can work.
So when the development team turns up for Day 1 of the development part of the project - lo and behold - most of the factors over which their experience should certainly have been used as the final decision criteria (e.g., toolset, along with team composition, architecture, etc) have well and truly been bedded down. Perhaps it's these sort of situations that led to Weinberg's Ten Percent Promise theory in his text, The Secrets of Consulting, whereby you should never promise to provide more than a 10% improvement to the situation you're addressing?
As a proud consultant who is a little miffed of seeing so many opportunities for improvement wasted because of decisions made prior to my arrival on the project... Grrrrrr!
Posted by Andy Marks at 02:54 PM | Comments (0)