CTRL-ALT-DEL Back to the Future, Part II 4th edition A revised speculative article (c) 1989-1993 H. Lee Siddons, Jr. Fathom April, 2000 [non-interactive, standard text version] (Please double-click voice annotation below for embedded voice version of this article) Volume 10, Number 4 Preface: "The future ain't what it used to be..." It's been almost ten years since the "Fathom" was first published, and as many of you are well aware, quite a lot has happened to the computer industry as a whole, Computer Associate's product line and the evolving database/programming environments. We are now on the verge of a new century, and many speculative articles and commentaries have been published about the industry and where we are heading as traditional programmers, class client/server/interface engineers and multimedia authors. This article is a first in an evolutionary series which will explore these issues by looking back at the changes in the last ten years: a retrospective view if you will of where we've been. 1990-1995 Ten years ago, in April of 1990, most of you remember Summer '87 marked the beginning of the end of the traditional DOS, character-based flagship products from the defunct company that created Clipper, Nantucket, with yet another major upgrade, Clipper 5.0 released in 1990 for the MS-DOS environment, that archiac OS which though still used today in many so-called low-end applications, has followed the same ill- fated destiny of CP/M. That release finally provided a solution to the memory management problem, VM (Virtual Memory and dynamic overlays), replacing the further need if desired for any overlays! And lest we forget how 5.0 and the VM system had major bugs which were fixed in the 5.01 maintenance release. Memory was only one of the problems all software faced as it was developed in the early nineties. Some of us back then were already pushing those limits of the current hardware. (I mean, who these days uses a 486DX/50 ? Sssllllooowwwnesss). Then, to the surprise of some, Nantucket was purchased by Computer Associates in 1992, the largest mainframe software company in the world. We have since seen CA/"Classic" Clipper mature in 6.0 and now the current 7.0 release. The core language, originally based on the archaic dBASE dialect (also abandoned by Borland in favor of its own Object dBase flavor), evolved and began to diversify as the then Nantucket moved it into a "C"-like OOPS direction. Not only were we able to use alternative RDD (replaceable database drivers), such as Paradox, SQL server (Microsoft/Sybase) and Oracle (Biton), but we could ourselves augment the list of Clipper language reserved words and functions with UDCs, the User Definable Commands in the CH preprocessor files and manifest contants via the CA/Clipper 5.X/6.0 language API. The first object classes packaged with Clipper 5.01, included the Get system, Tbrowse and Error system. CA-Clipper 5.2 added new commands, more RDD's for FoxPro, Paradox and dBASE IV files. The network, CUA/SAA, GUI and mouse object classes came in the 1994 release. These were ambitious products for us, as we had to rethink our programming nomenclature by adapting to new concepts: recall all those articles and user-group talks we had back then on object-orientation, cross-language binding, code blocks, message passing, classes, encapsulation, inheritance, methods and "arrhhgg" polymorphisms??). There were two camps of thought: the left-brained traditionalists who were comfortable with their proven and familiar data and procedures ('a la Summer '87), and the more adventuresome right-brainers (5.X) and CA-VO, who saw elegance and creativity in WYSIWYG object-orientation. The infamous battle between the Clipperheads and GUIes had begun. Some of us were initially intimidated by these "new" terms and resisted these concepts as they were debated and discussed in the CA-Clipper forum (CIS) and through previous articles here. We also learned how to embed and/or call early standard SQL into our programs easily by just using the CONNECT/SELECT statement. This enabled the manipulation of large pieces of data more efficiently as well as across other platforms and databases, collating mainframe DB 2 data as well as client-server based SQL data. New data types including nils, code-blocks, formula-based, large binary (incuding graphics, hypertext, AVI), object-class and even user-defined have since come into being. Remember when the CLIPPER C.O.R.E. group, a standards committee formed in '94 which later disbanded? It helped pave the way to avoid any confusion as to the diverse directions the Clipper core language was heading New OOPS dialects could be created by anyone, so Clipper standard BNF's and guidelines were adopted that went beyond the Clipper guidelines published in the XBase News. This was brought about because several third party developers were providing redundant variations, of new command sets and early dBASE IV/Object dBase (ie: Matt Whelan Enterprises' the Professional Compiler Kit for dBASE IV), FoxPro, etc. dialects that were in some cases incompatible with one another. Neat features like an incremental compiler/linker (RT- Link) and full screen symbolic debugger, including the API and later the CA/OOPS SDK for third party developers only wetted our appetite for more...this was only the beginning. Besides the CA-Clipper series, in 1992 and late '93, Computer Associates introduced the first of a set of new products, CA-VisualObjects (a.k.a. ASPEN), with other related components like code and report generators that would take advantage of both object- oriented programming concepts while making some available under new operating environments. At the 1992 TechniCon fall session held in New York, Computer Associates unveiled its client-server stradegy which included CA-Clipper 5.2, dBFAST 2.0 for Windows 3.1 and at the 1993 fall TechniCon in Orlando CA-Visual Objects 1.0 for Windows 3.1, NT and OS/2. The early 1990's saw this gradual change from the Summer '87 scenario to the earlier known research project called Nantucket Future Technology's ASPEN, an ADP (Adaptive Platform Designed) and GUI-based, object-oriented, hyperfile format most of us enjoy today. But competition of sophisticated Windows based DBMS's heated in early 1993 with the release of major players, Microsoft and Borland. Microsoft's FoxPro for Windows 2.5 and it's innovative Access 1.X, notwithstanding Paradox for Windows from Borland, placed dubious doubts in CA's lauded, albiet late entry. In response, several third party developers even introduced their own serious versions of Clipper for Windows. But some developers found it neccessary to abandon Clipper for other Windows-based development systems because of demand. At that time, VO meant "Vapor" Objects to those who doubted it would be released in time to succeed in the marketplace. Computer Associate's CA-VO for Clipper under Windows 3.1, much of which was written in assembly made it one of the fastest compilers in the world. But it was still under the limiting umbrella of pre-7.0 MS-DOS. Visual Objects for Clipper under Windows NT or WIN32 changed all that: the beginnings of our current hybrid version has it's roots here! We now could write true 32-bit applications that could run not only on Intel-based, but also on RISC-based and SMP (multiprocessor)-based machines. 1995-1999 We could still use much of the same code you had written earlier, as long as you followed standards. But if we wanted to go beyond just thunking 16-bit apps and really wanted to take advantage of the new data structures, multi-threading, graphics, communications and MCI (Media Control Interface drivers) using multimedia technology, all of your programs written for Windows NT (New Technology) would have to be redesigned. The security and new file formats introduced in Windows NT's NTFS (New Technology File System), provided a mainframe-like alternative to the archiac DOS-based FAT table structure. Taking the best features of FAT and OS/2's HPFS, NTFS added security and fault-tolerance and we now could use 254 characters for file names. Routines for early DDE (Dynamic Data Exchange), DDEML, ODBC (Open Database Connectivity), OLE 2.0 (Object Linking and Embedding), techniques for storing and accessing graphics (earlier TIF and JPEG), as well as TrueType font I/O management were the subject of technical discussions back then. VO/CLIPPER writers and developers debated and provided solutions to such issues as to the aesthetic use of fonts like Arial verses System and how best to adjust your program's screens and fonts proportional to the myriad of screen/display resolutions; how best to manage and maintain your CA-VO configuration/registration databases; how to invoke calls to OLE 2.0 to use Word for Windows NT or WordPerfect for Windows NT as the VO editor for memos and/or text data objects; CUA/SAA compliance for non-corporate based applications; to even debating the storage requirements as well as the legalities of cut and pasting real time broadcast TV pictures into CA-VO graphic AVI objects via TV/digital boards. Since the standard interface of pull-down windows, dialog boxes, radio buttons and bit-mapped graphics provided a new paradigm, in subsequent VO releases, more emphasis was placed on RDD/ODBC features and performance as well as new multi-tasking/threading optimization of index algorithms, hypermemo text techniques and device drivers that would take advantage of the latest hardware advances. CA-VO for Windows 4.0 (Chicago) and NT 4.0 (CAIRO) debuted, which realized newly designed object- oriented device independent applications with the help of many CASE (Computer Assisted Systems Engineering) tools hitting the market back in 1994. The plan for a mainframe version, CA-Clipper Unicenter- Mainframe, was later shelved since mainframes have gradually been phased out through downsizing, used instead as high-level platforms in favor of mega-data repositories. TODAY & TOMORROW Hardware advances helped to complement the evolution of the current VO/CHAT (Clipper Hybrid Aspen Toolset). Since mega memory chips are not as cost prohibitive as they were ten years ago, the newly released Windows NT 5.0/OS/2 API combo now can run comfortably under 16 MEGS of RAM, and if that's not enough many of us will have new 250 MHZ 64-bit P7 systems equipped with enough virtual dynamic laser storage to handle any overhead. Next year, or as many of us like to say, "next century", a new prototype multi-media HDTV/computer will be on the market from AT&T/Zenith, which replaces the Smartphone series. This piece of advanced hardware finally promises the first true connection between the hybrid ATM/FDDI/ISDN-based digital voice/data communications protocol and the digital fiber optics television network. It will run under several operating systems including Windows NT 5.0. It comes standard with an internal FAX/satellite/common carrier management system, digital VR (voice recognition), 32 megs of internal RAM for DVI/HDGA as well as a 20 GB RW laser hard drive that accesses all CDRW/CDV and HDCD formats. Industry experts also expect external analog as well as digital interface capability with many of the new popular smart home LANS. The system effectively replaces the traditional phone, computer, TV and stereo and merges them into one integrated component. All this fits in a briefcase for under $10,000. I just might want to get one of those. . . So what does this mean for developers? Well, when the CA-Clipper/VO product line is finally available via a modular- code static chip (which Computer Associates has in beta now), only our imagination and creativity driven by consumer demand will dictate the implications of this. *** End of Article *** Note: The first version of this appeared in April '89 in the Sunshine Clipper, the second incarntation in the Nantucket News in March/April 1990 issue. The last one appeared in the July 1991 edition of the Fathom, the Orlando Clipper User Group newsletter. This is the latest and fourth incarnation for which major changes were incurred due to the CA purchase and other changes in current reality. -= This file was brought to you via ClipperNet =- ñmax