Arthur Fuller
artful at rogers.com
Thu Aug 18 11:21:00 CDT 2005
About a decade ago I headed a development team for an insurance company. Half the job was to take some mainframe programmers and acquaint them with PC development tools and best practices. XP was not a big thing then, but I improvised an approximation of it at that time. I started with 2 programmers, and had them sit with me as I coded. They were learning lots of things at once -- syntax, editor keystrokes, shortcuts, etc. At first I was moving so quickly that they couldn't even see what I was doing (particularly in the editor itself), but a few explanations and we were over that hump. The hardest thing to teach them was "1 proc does exactly one thing"... No, upon reflection, the hardest thing to teach them was "avoid coding until the last possible moment." The first two programmers then took on a programmer each and repeated the "semester", as it were, and then those two took on a programmer each. By the end of the project, we had a team of capable programmers, and we stuck to the rule of "two programmers per keyboard" all the way through the project. It worked brilliantly -- about equivalent to debugging while writing the code. The second programmer might watch the first start a recursive routine and say, "Hey, wait, let's break it out into two routines and avoid the recursion problems." They might kick it around for a bit before choosing one or the other approach, and also document why they elected that choice. I am _way_ big on BDUF! Avoid coding until the last possible moment -- that's my motto. I started out as a cowboy coder, but no more. I spend a _lot_ of time db-design tools such as DeZign, and I don't even begin creating sample data until I've been through about 3 revisions of the given DB design. Then I write use-cases describing each function the software must perform, using English not pseudo-code -- in fact, I have learned to avoid any specific references to tables, and instead to write such statements as, "When the user selects a product and quantity she wishes to buy, place a 5-minute hold on that quantity. When the user's credit card payment is authorized, deplete inventory by the specified quantity and remove the hold." Once the DB design is finalized, and the purpose of each table, view, sproc and UDF described, then even a relatively new hire can take the spec described above and implement it easily. In the example I gave, there are the following steps: 1. Obtain the product ID and desired quantity 2. Check to see said quantity exists (we'll sidestep the issue of back-orders for now). 3. Place a hold on said quantity, with a timer mechanism of some sort (SQL Server is great at this, much better than an Access timer). 4. Obtain the credit card info, fire it off and check the return status. 5. If the payment went through, deplete the quantity available. 6. Remove the hold. Each of these steps can be reduced to exactly one procedure. You (as manager) could even assign each step to a different programmer. So long as the DB design is solid, there should be no issues. I haven't read either book you mentioned, but I'll put them on my list of things to read. Arthur -----Original Message----- From: accessd-bounces at databaseadvisors.com [mailto:accessd-bounces at databaseadvisors.com] On Behalf Of Shamil Salakhetdinov Sent: August 18, 2005 5:50 AM To: !DBA-MAIN Subject: [AccessD] OT(?): Big Design Up Front (BDUF) vs. XP any comments,ideas, experience? Hi All, I've got this this today from JoelOnSoftware subscription list: <<<<<< "As I worked through the screens that would be needed to allow either party to initiate the process, I realized that Aardvark would be just as useful, and radically simpler, if the helper was required to start the whole process. Making this change in the spec took an hour or two. If we had made this change in code, it would have added weeks to the schedule. I can't tell you how strongly I believe in Big Design Up Front, which the proponents of Extreme Programming consider anathema. I have consistently saved time and made better products by using BDUF and I'm proud to use it, no matter what the XP fanatics claim. They're just wrong on this point and I can't be any clearer than that." - From my latest article: The Project Aardvark Spec http://www.joelonsoftware.com/articles/AardvarkSpec.html >>>>> No comments. It happened that I'm currently reading: "Microsoft Object Thinking" by David West ISBN:0735619654 Microsoft Press C 2004 and "Extreme Programming Adventures in C# " by Ron Jeffries ISBN:0735619492 Microsoft Press C 2004 These two books are from Microsoft Press and they have a lot of useful information on eXterme Programming, Unit Testing and Classical(Behavioral) vs. currently existing in most implementations "real life" UML-based(RUP) OOD&P. When MS Press publishes books on such more computer-science than used in real life projects subjects then they are very probably "cultivating the ground" for the soon to become true "dreams". (I remember I watched a movie "Microsoft, Year 2004"(form MS of course) somewhere in year 1995 or so - and as I see now their by that time science-fiction is now real-life, exactly in year 2004-2005. I must say I impressed with both books (and there are just a few technical books I liked because most of them are just "chewing" MSDN or other docs without almost any new ideas). The more I read them the more I like them and I see a lot of sense in what is written in them and I see a lot of my own ideas and experience are there too. I'm glad my thinking is in the mainstream (of course what they say is much more elaborated and thought through and based on their own rich experience). So my guess/thinking is that XP, Unit Testing and Behavioral (Extreme) OOD&P are becoming mainstream for real-life development of the projects of any size - in fact as authors of these books state (based on their experience) that the stuff they are writing about is the "only" agile way to solve the challenges of nowadays customers and projects requests. And they are not fanatics I think - they base their writing on deep analysis of all the previous 50 years experience in software development and not only in software - "Microsoft Object Thinking" is more philosophical than technical book and it has quotes from Plato (year 400 B.C.) - these quotes are used to explain how to "attack" complicated/vague projects' requirements... What do you think about the subject and related issues? Do we need Access/VBA Unit Testing added here - http://opensourcetesting.org/unit_misc.php (Open source tools for software testing professionals) - it can be done - does it make sense? Shamil -- AccessD mailing list AccessD at databaseadvisors.com http://databaseadvisors.com/mailman/listinfo/accessd Website: http://www.databaseadvisors.com