A public forum for discussing the design of software, from the user interface to the code architecture. Now closed.
For a long time now, we've been arguing that the architecture our systems are built on is quite poor. Finally we are starting to get some buy in from key management and they might finally agree to bring in a 3rd party to analyze the architecture (with us saying it's crap, the consultants saying it's great, we figured we needed a tie breaker).
Problem is, can't say I know exactly what formal artifacts and tasks to place in our statement of work. I've found very little about actual reports/methodologies of post-develoment analysis. Our concern is that consultants were real astronauts and built systems that only they'll even be able to maintain. Maintenance and knowledge transfer are problems , performance continues to be a problem. Best practices weren't followed...
Anyone have anything useful? What formal artifact would a detailed system/architecture analysis result in?
Monday, December 05, 2005
If you really mean architecture as opposed to design, then I can give a good recommendation to ATAM ( http://www.sei.cmu.edu/publications/documents/00.reports/00tr004.html ).
Thursday, December 08, 2005
Ugh. Let me quote from Matt's link to describe the problem, "architectures are complex and involve many design tradeoffs."
The formal artifact you need is unique to your business needs. For example, here are some of the things you will need to consider...
Reliability and uptime.
Data redundancy, backup and restore.
Networking and Bandwidth.
Ongoing maintenance and support costs.
Physical space (or co-locate?) and physical security.
...and you will need to decide how important each of these criteria are. Take data for example...
If you're a small business with data that's not very valuable (you simply have to pay lots of money to type it back in), you can get away with a simple DVD-ROM backup server in a closet.
If you're a small business with very valuable data - and financial consequences if you loose it or it gets stolen, you may want to look into remote backup solutions.
If you have a LOT of data, you may want to look at maintaining RAID servers, SANs, NAS, the works.
And so on. Essentially what you're trying to do is assemble set of documents that asks all of these questions about the business and records the answers in such a way that you can make good planning decisions.
Is it cheaper to buy 20 individual antivirus packages or get the enterprise license that requires a server? The answer is somewhat driven by the question "How many computers do you need to protect?"
Friday, December 09, 2005
Heh, there's a 50/25/25 chance here - 50% chance that the third party will try to claim both existing sides are idiots and take over, 25% chance that they'll agree with you, and 25% chance that they'll agree with the people who you think are idiots.
As you've already established that management doesn't completely respect your opinion in this matter, you've got a 25% chance of getting a chance at making some changes.
Ok, so I made up numbers - still, getting a third party in as a tie breaker indicates that there's some unresolved issues that aren't going away even if you get the result you like.
The real fun will be when the manager who hired the consultants finds that his pets lost, and has to cover his own ass.
Dear shiva, jesus and zeus, I am so glad I am not you.
Monday, December 12, 2005
This topic is archived. No further replies will be accepted.Other recent topics
Powered by FogBugz