home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.software-eng
- Path: sparky!uunet!zaphod.mps.ohio-state.edu!wupost!udel!rochester!rocksanne!news
- From: kirby@xerox.com (Mike Kirby)
- Subject: Re: Appeal to SEI (was: Debugging the proc
- Message-ID: <1992Dec28.220147.11493@spectrum.xerox.com>
- Sender: news@spectrum.xerox.com
- Reply-To: kirby.roch803@xerox.com
- Organization: Xerox Corporation, Webster NY
- References: <1992Dec28.193911.20735@asl.dl.nec.com>
- Date: Mon, 28 Dec 1992 22:01:47 GMT
- Lines: 55
-
- In article 20735@asl.dl.nec.com, terry@asl.dl.nec.com writes:
- >Hi folks,
- >
- >In article <1992Dec28.113148.15707@sei.cmu.edu> mcp@sei.cmu.edu
- >(Mark Paulk) writes:
- >
- >> What we intend the CMM to say is that an organization has to have a good
- >> metrics program in place to achieve level 4... We do not say that an
- >> organization cannot, or should not, have metrics in place much earlier...
- >> The belief that the SEI requires that "you have to be level 4 to do
- >> measurement" is a fairly widespread misconception that we have tried
- >> repeatedly to correct...
- >
- >From the perspective of a critic, I would definitely concur that the SEI
- >concept of process progress has always had a strong metrics flavor from
- >the get-go. It shows up clearly in level 2, let alone 3 and 4. Indeed,
- >one of my complaints was that it unclear what a level 2 organization as
- >defined in early SEI materials would DO with their metrics it was clearly
- >supposed to collect.
-
- KEY KEY point. What do you do when you collect process metrics from a level 2
- organization. This means that your metrics are management metrics, and there
- are no engineering metrics because you have no engineering process. Now that
- we have our measurements, at some point we must decide that our process is
- bad based on our measurements and then change the process to bring our
- metrics back to a reasonable level. However, we have no way of determining
- how this will affect the engineering non-process. If we had a process,
- we could measure it, and we could measure the effect of management process change
- to ensure that it isn't causing any undo problems. Of course this doesn't even
- begin to talk about what are the "right" metrics and what are the "correct"
- values.
-
- It is my belief that product metrics are much more effective in non-process
- shops. It is fairly easy to measure something like defects/kloc or complexity,
- and order additional testing or longer development time based on those numbers.
-
-
- >
- >
- > Cheers,
- > Terry Bollinger
- >
- >
- >+--------------------------------------------+-------------------------------+
- >| Terry Bollinger | Phone: 214-518-3538 |
- >| Advanced Switching Laboratory, NEC America | Fax: 214-518-3499 |
- >| 1525 Walnut Hill Lane, Irving, Texas 75038 | Email: terry@asl.dl.nec.com |
- >+--------------------------------------------+-------------------------------+
-
-
-
- Mike Kirby
- Xerox Corp
- E-mail: kirby.roch803@xerox.com
-
-