I think the problem lies in the mindset of the programmer. Every budding programmer, when asked, will readily recite to you the tenets of OOP. But do they really understand them enough to *use* them ? No, I don't see it happening at all. Writing performant code requires you to know about ten different ways of accomplishing the same thing, each of which is suited to a particular scenario, so that you can decide which one is best in this case. How many programmers can think of one method (let alone ten), without having to resort to Google ? Writing secure code requires you to think pessimistically - "All input is evil unless proven otherwise.", for instance. How many programmers validate every single piece of input they receive, as a principle ?
For most programmers, code reusability is being able to copy and paste the code everywhere with minor modifications, but that's not actually reusability. The problem is also that procedural programming (as opposed to Object Oriented) is too easy and doesn't require you to think about the possibilities beforehand. It allows you to dive right in and apply the patchwork as you go. Secondly, the problem is in the education. We stress inordinately upon the theory of OOP, rather than the implementation. I think OOP is not something to learn... it is a skill to inculcate. And it takes sustained effort and strict personal standards. Most IT certification exams expect a candidate to select a correct option in a multiple choice format, instead of requiring them to come up with some code. The problem also lies in the mindset of the Management. Tight delivery schedules and strained resource availability further reinforce a focus on getting the job done, not on how it was done or how scalable and performant it was. If (or rather "When") problems are encountered later, quickfixes are the best solution to the matter, rather than a complete redesign to match changed requirements. You can get an idea of the situation from the fact that most of my colleagues regard me as eccentric because I am the only one who will not compromise on the quality of my own or my team's code. You ask why we're still at the same level as years ago. I think it's because the creation of software is driven by business considerations and not the sole pleasure of creating software. The reward is not in the work itself but in the (material) gains associated with the completed product. Maybe that's why coders are not artists, but just labourers. *rant over* ;-) On Oct 31, 1:48 am, flatfilehater <[EMAIL PROTECTED]> wrote: > So I am not the last survivor of the Osborne 'laptop' age! > > I am a coding addict - I still get a kick out of having things work > and seeing others work with it. > > the root of my gripe is actually stuff like php and mysql which most > isp's over here (south africa) use. > > when i attacked these things, it was a step backwards (probably many > steps). it was like dealing with random access records from flat > files (but with built in filters and indexing). I thought I had died > and this was hell. > > i am getting into xml (slowly) and am searching for a way to deal > decently with many to many's. > > any ideas out there? > > On Oct 21, 8:16 pm, Glenn <[EMAIL PROTECTED]> wrote: > > > > > Ah, someone else that remembers that era!! > > > I guess there are still enough of us out here that still like to write > > code. Unfortunately, most of the "code generators" out there don't do a > > very good job or don't have the ability to generate all the possibilities > > that we need. Therefore, we use them to get close and then fill in the > > blanks with actual coding. > > > By doing actual coding I think we can achieve the level of re-use if we code > > it properly. After a bit of coding, we find that we collect a large enough > > library that we can start re-using what we have already written. However, > > many students out there that frequent this group only accomplish what they > > can with the code generators (a.k.a., wizards) and don't go beyond that > > stage. > > > I think it will be long after we're out of here that someone will be able to > > sit down at a computer, tell it what type of program it wants and have the > > computer create it for them. However, that only seems to exist in science > > fiction right now. > > > ...Glenn > > > On Tue, Oct 21, 2008 at 1:16 PM, flatfilehater <[EMAIL PROTECTED]> wrote: > > > > as reply to mails, i am from pre-ms generation. > > > > when ms arrived the things had to be ibm compatible - it worked by > > > creating a relatively common denominator (lowest maybe) for making > > > mass software feasible (lotus 123 was the spreadsheet of the era but > > > visicalc was the first; ashton tate's dbase became the database entry > > > level later). > > > > the paradigm shifted with borland turbo pascall - libraries of code > > > were applied to cut down repetitive coding (i think borland was the > > > first to use the term window - the mouse arrived with ventura > > > publishing). > > > > this shift was a breath of fresh air and we all said that we were well > > > on the way to a new generation which would automate even further and > > > almost interpret things into plain english. > > > > now about 25 years on, we are still tick-ticking in code via a babel > > > of languages - analyse the repetitive nature of the info being sought > > > in the group and it seems that most are reinventing a wheel. > > > > of course i am getting old and cranky but not without reasonable cause > > > - the very attainable things and logical progress have not happened. > > > > time for coffee and a smoke!- Hide quoted text - > > - Show quoted text -
