Delphi: Old times - Opinion

Old times

A review of the old times

It is said that any technological revolution requires 30 years to become a definitive and daily form of the human life. It has been this way with the television and the telephone. Computing in some of its aspects (as a revolution) seems to be arriving at its thirty years and thus we can assume that it's already taking its definitive form.

Since the late eighties, in desktop computing, there hasn't appeared a truly new (original) product that is important outside its operative surroundings, the last one was Access, Microsoft bought it from a French company as I recall, and if we are rigorous it's a sort of visual evolution of the database systems (xBase, for example). We have about 10 years seeing how existing programs only change their version. Even new manufacturers only make new versions of already existing programs. Think about this a little, Star Office and MS-Office come from WordStar, Visicalc (or Multiplan) in the 80s. Programming languages are not new, just evolutions of BASIC, Pascal and C. Internet browsers are as new as the www (which is not really very new). Checking my computer I didn't find any software really original that wasn't dependent of the operating system (FindOrphan, StartupCop for example).

The previous introduction comes to collation because I believe Delphi is arriving at its definitive form as an evolution of Pascal. Let's take a look at the past. I don't intend to tell again what many of use already know, I just want to make some curious comments about history. From now on I'll be referring to Delphi as an evolution of Pascal.

A crucial historical moment. The mid-eighties.

Lotus 1-2-3, CP/M, PC-AT and the most common language (at the level) of microcomputers: BASIC.

Somewhere in the world, in a cold night, a miserable programmer on his knees finished a prayer like this "... and get us rid of BASIC, amen." Then Turbo Pascal appeared. This is the kind things that makes me doubt about the nonexistence of God.

Delphi arises as an evolution of Pascal and Pascal is a prototype (like Ada, APL and others) of what has been called a B&D language ("bondage-and-discipline"). The image of an English woman dressed in black with a whip and chains, determined not allow the possibility of leaving certain rules ('right programming') is very clear to understand for those who went thru "BASIC" in the eighties. Turbo Pascal is responsible for the first valid attempt to take out two labels about Pascal: B&D Language and "toy language".

Toy Language

As toy languages we understand those languages of educational purpose designed as a test of some theory in computing science. Pascal, as Niklaus Wirth defined it in 1967, was that. In fact, for a long time many people thought that taking Pascal to a general-purpose language was a bad idea. Among them was Brian Kernighan.

In 1981, Brian Kernighan wrote a paper titled "Why Pascal is Not My Favorite Programming Language". As far as I know you can look for it in "Comparing and Assessing Programming Languages" of Alan Feuer and Narain Gehani (Prentice-Hall, 1984). There he manifested some interesting things. He was right in at least one: language extensions end with the portability. The fact is that, returning to present times, if it weren't for the limited portability towards Kylix, Delphi isn't portable (yet). Anyway, one of problems he saw were those of the types "string" (an extension that isn't part of standard Pascal), static and global variables, and other evils that OOP has tried to eliminate with quite success.

I never programmed in standard Pascal, my first Pascal was Turbo Pascal 2.xx in CP/M, that perhaps was a toy language as Kernighan considered it, but it could get everything that CP/M was able to give (Would it mean that CP/M was a toy?).

Nowadays many Delphi programmers come from C and BASIC in their more evolved versions (Visual C++, Visual Basic) and I don't believe that they think of Delphi as something inadequate for general-purpose programming. Clearly, Delphi is more than a superset of Pascal.

The publication of this material is allowed by any means from anyone as long as this material is not modified and the original source is mentioned.

Second part

Soon we'll publish the second part of this article.