Wolfram Mathematica Online Integrator
-
Member 4593559 wrote:
I'd be interested to know what automated tools you are using. Currently we are using the PSP Dashboard, which is an Open source project
It is a combination of a VS line counter plugin I found and PSP Dashboard/Excel. Not exactly the prettiest, but it does okay since my shop is small. Not sure about yours. The only thing that still bugs me after being introduced to PSP in 2003 and doing my Master's Thesis on it, is the notion of the defect tracking. When talking about how many errors per 1k/loc, are we talking about silly little errors, or logical runtime errors that are found? Never been 100% clear on that. I've just filed it as something that is decided by the user of PSP. I lean towards the logical errors, because counting syntax errors based on a missing semicolon or curly brace is rather pointless. Counting syntax errors just shows how crappy of a typer someone can be, but the logical errors are what is really more important.
Member 4593559 wrote:
Of the data I have collected using PSP, none of it seems to have any intrinsic value when it comes to predicting how long my next project will take or how big it will be, and it just seems like an awful lot of red tape for little or no gain.
I know what you mean. Most of my client work I do has been fairly well predictable with and without PSP.
Member 4593559 wrote:
The spreadsheets that come with the course notes for PSP and TSP are laughable
I agree. I cannot begin to say how many times I've taken an Excel spreadsheet and tweak it around.
"The clue train passed his station without stopping." - John Simmons / outlaw programmer "Real programmers just throw a bunch of 1s and 0s at the computer to see what sticks" - Pete O'Hanlon "Not only do you continue to babble nonsense, you can't even correctly remember the nonsense you babbled just minutes ago." - Rob Graham
Lol. I guess that there isn't a seriously good tool out there yet for much larger scale operations yet then. Where I work we have Teams using TSP and the associated spreadsheet in conjunction with the Dashboard, and a few softies have attempted to craft line counters...I'm not sure but I guess they must have standardized on one. We do hear a lot of complaints about it from the teams, but the company has paid a lot of money to train its engineers in the principles of PSP and TSP, so I guess they are going to persist with it. Fortunately for me, I am currently working on a lot of Custom projects as a team of one, so I have elected to just stick to the basic principals of Software Dev. I remember when taking the course, we were encouraged to disable intellisense, not hit the compile button until reviewing the code, and to record even the smallest defects, such as missing semicolons. To me, that just seems like a way to create false impressions of high defect rates on projects. Modern IDEs and compilers have moved on so much since when PSP was originally developed, that to not use tools like intellisense that were created to aid developers is just backward. Its true, that since the course they have had someone visit from the SEI, and tell us that we only need to take what we need from the principles of PSP, but this should have been made clear on the course. my curiosity was piqued with the talk of Agile methods, but I'm beginning to get more and more jaded with all these different approaches and Magic bullets that promise quicker software turnaround with fewer bugs. I like the idea of softies being craftsmen, and the trade taking a while to learn to make code truly exceptional quality code, as opposed to large scale Flat pack, knock 'em out quick approaches. A Tad romanticized, perhaps, but an altogether more attractive proposition ;)
-
Lol. I guess that there isn't a seriously good tool out there yet for much larger scale operations yet then. Where I work we have Teams using TSP and the associated spreadsheet in conjunction with the Dashboard, and a few softies have attempted to craft line counters...I'm not sure but I guess they must have standardized on one. We do hear a lot of complaints about it from the teams, but the company has paid a lot of money to train its engineers in the principles of PSP and TSP, so I guess they are going to persist with it. Fortunately for me, I am currently working on a lot of Custom projects as a team of one, so I have elected to just stick to the basic principals of Software Dev. I remember when taking the course, we were encouraged to disable intellisense, not hit the compile button until reviewing the code, and to record even the smallest defects, such as missing semicolons. To me, that just seems like a way to create false impressions of high defect rates on projects. Modern IDEs and compilers have moved on so much since when PSP was originally developed, that to not use tools like intellisense that were created to aid developers is just backward. Its true, that since the course they have had someone visit from the SEI, and tell us that we only need to take what we need from the principles of PSP, but this should have been made clear on the course. my curiosity was piqued with the talk of Agile methods, but I'm beginning to get more and more jaded with all these different approaches and Magic bullets that promise quicker software turnaround with fewer bugs. I like the idea of softies being craftsmen, and the trade taking a while to learn to make code truly exceptional quality code, as opposed to large scale Flat pack, knock 'em out quick approaches. A Tad romanticized, perhaps, but an altogether more attractive proposition ;)
Member 4593559 wrote:
we were encouraged to disable intellisense, not hit the compile button until reviewing the code, and to record even the smallest defects, such as missing semicolons.
Forget disabling intellisense. Like I had mentioned, counting missing semicolons is kind of pointless since it does not affect the end product. You don't see software having updates in the binary because someone forgot a semicolon. The compiler catches that before the customer/client even sees the deliverables.
Member 4593559 wrote:
To me, that just seems like a way to create false impressions of high defect rates on projects
I agree. Just like any other kind of metric, you can get someone who misconstrues it or fudges it around to look better or worse that it really is.
"The clue train passed his station without stopping." - John Simmons / outlaw programmer "Real programmers just throw a bunch of 1s and 0s at the computer to see what sticks" - Pete O'Hanlon "Not only do you continue to babble nonsense, you can't even correctly remember the nonsense you babbled just minutes ago." - Rob Graham
-
It was broken by Mustafa first... You can't blame him. ;)