Clipper still exists?
Clipper was a language much used in the 80s and until much of the 90s. Especially in some countries like Brazil. In fact many administrative systems still used today were made in it. But is it still possible to develop modern software using it?
2 answers
Fatal Fate
The Clipper, that product created by Nantucket and purchased by Computer Associates, was discontinued in development around 1996. It was still marketed for a long time and until a short time it was still possible to buy it by a third party who made a deal with CA.
Clipper had its fatal fate when CA developed the CA-Visual Objects which was the most modern, object-oriented, slightly lower-level Clipper with a complete IDE. The product was poorly developed despite some very interesting ideas and was a complete failure. Part of the problem is that Clipper programmers could not understand some differences from the new language that was similar to the old one but not completely compatible (which also contributed to the failure).
There they did not have an interesting new product and the old one was abandoned. They did not believe that it gave to improve the old and that it still had people interested in him. It helped that the product did not make much success in important markets.
But the product was good, had interesting innovations and did some things that were only found in successful languages many years later.
Some clones of the language that was already a clone of the dBase simplified database language emerged.
A hope arises
The great success, if we can call it that, was with the Harbour , a version open source compatible with Clipper. Developed from 1999.
In addition to giving total freedom to developers for not having a company but a community because of it, it gained an automatic modernization since it abandoned a 16-bit C compiler created in the 80s by any other modern 32-bit C compiler, and later 64-bit. Running virtually on any platform that has a C compiler (some with a bit of effort). In a way Harbour codes are transformed into C codes generating executables always.
In addition to this extensions to the language were created while maintaining compatibility with legacy codes and many new libraries were gradually developed, such as GUI ( Qt is Just example), sockets, compression, new data structures and algorithms, encryption, images, Unicode, XML, clients for various relational databases or other technologies, better access to operating system, best ways to access your own database system(dbf), multi-thread and much more.
The language has gone on to have extensions that allow object orientation in a way very similar to that adopted by other dynamic typing languages, gained pointers( which are actually references), structures of hash (which are not actually hashes in fact, it is, Harbour people are bad at naming), for each , switch , better literals, dynamic loading of code, compiled or not, and constructs that avoid using some legacy features considered bad as well as easier integration with C code.
Some improvements have also been added to the preprocessor which is absurdly better than the existing one for C (it's not having a preprocessor that is bad in C, it's having a bad one).
In addition to this the compiler improved some checks and the entire infrastructure, in addition from the compilation, it was modernized.
Conclusion
Does not mean that we will see new programmers using a language that owes almost nothing to the most modern languages but those who have legacy code or are well adapted to technology can still continue to evolve their softwares using something that would surprise programmers of PHP, Python, Ruby, JavaScript, etc.
I just don't understand why there are still programmers who still keep using Clipper original, sometimes an old version of the 80s.
So despite having people who still use it, you can no longer use it directly (16 bits!!!). What gives is to use the modernized version. There is a better way to use the same technology. Not only Harbour, there are other alternatives to Clipper (I do not know them well) such as xHarbour (a semi-standard fork of Harbour), xbasse UI, FlagShip, FoxPro (abandoned by Microsoft), AdvPL (proprietary language used in ERP Protheus Da Totvs).
In the 90s I made some applications in DBaseIII-Plus, dBase IV and Clipper, preferring Clipper to dBase (s). But with the appearance of Microsoft Windows and Clipper running only in DOS environment, I opted for FoxPro, but did not like it and abandoned the Clipper.
However, three years ago I discovered Harbour and liked the way it is structured and its potential, including the portability of applications for *NIX environments (BSD, Mac, Linux, etc.) in addition to traditional systems Microsoft operations, links to relational databases beyond traditional DBF files, creation and use of guides, etc..
After starting the development of a management application for an association, I soon opted to switch from Harbour to Python, for several reasons:
- development in Harbour was being slower than if it were done in Python;
- I started to notice that it would be harder to keep the code in Harbour than in Python;
- Although the Harbour was already stable, there were still some uncertainties in the direction the Harbour would take;
- The Python community is larger and offers more support possibilities than the Harbour community.
To this day I do not regret the choice I made.
Note that Harbour is "open source" while xHarbour is proprietary.