From mboxrd@z Thu Jan 1 00:00:00 1970 X-Spam-Checker-Version: SpamAssassin 3.4.5-pre1 (2020-06-20) on ip-172-31-74-118.ec2.internal X-Spam-Level: X-Spam-Status: No, score=-1.9 required=3.0 tests=BAYES_00 autolearn=ham autolearn_force=no version=3.4.5-pre1 Date: 19 Aug 91 17:06:00 GMT From: bu.edu!inmet!stt@bloom-beacon.mit.edu Subject: Re: LOC estimates for Ada Command and C Message-ID: <20600117@inmet> List-Id: Re: LOC estimates for Ada Command and Control (from dennis@tfsg.UUCP) My personal favorite rule-of-thumb is 2.5-3 person-months/KLOC. 1 person-month for design/initial documentation 1 person-month for code/unit-test 0.5-1 person-month for integration/system-test I realize this isn't LOC/day, but months/KLOC, but I much prefer inverting the numbers since they suddenly become meaningful. Noone writes X LOC in a day, and then 2X LOC in two days, etc. However, people do work some number of months on something of some given size. So, for your 50KLOC, I would estimate 125 to 150 person months, or 10 to 15 person years. This number would come down if some of the 50KLOC is generated by tools, or reused from some earlier project, though integration testing can be longer in such an environment if the tools are not fully robust, or the reused code is not a perfect fit. Also, this estimate can come down dramatically if you are talking about a project involving just one or two expert programmers, with relatively well-defined requirements and modest documentation requirements. Unfortunately, this is rarely the case... Our experience with Ada has been that integration testing is shorter (because most interface errors have been caught earlier), while design is somewhat longer, than other languages. It isn't so much that design for Ada is harder, but that one tends to be inspired to do it more thoroughly. S. Tucker Taft stt@inmet.inmet.com