kenr@cruise.cc.rochester.edu (Kenneth C. Rich) (06/12/91)
Here is the promised Pick summary, 4 folks talking nearly simultaneously. >1. Tell me about Pick in general. Is it a Relational DBMS? I've been using PICK for about 7 years, developing applications for the College's Placement Office. I like it a lot. I did the initial software evaluation and chose PICK over more CICS based systems and an infant PC database market. Is PICK relational? Some would say no; detractors told me it wasn't before I chose the software. In a sense it is. Data can be spread out over several files, and a function in a file's dictionary can reference individual fields or the entire record from another file. > I hear it's an operating system and DBMS combined? > What architectures is it available for? Pick is similar to an RDBMS; it started out life as an operating system with extensive database capabilities built in. It is now available as a database product under Unix: Vmark, Ultimate, Prime and other companies offer Pick-under-Unix in a variety of flavours on several hardware platforms. PICK began its life as a operating system/DBMS (some still think it is an OS). Then inovators like Prime computer put PICK on top of their existing OS. PICK runs on top of Primos (Prime Operating system) as well as UNIX. There is even a new variant of PICK as an OS that runs on Intel PCs. (I use Prime's PICK variant, Prime INFORMATION.) PICK on top of another OS generally allows access to the OS's functions for the convenience of the user; some varients allow system level subroutines to be included in your PICK code. It's the best of both worlds. [from director of R&D at Vmark Software:] 1. PICK is an NFNF (non first normal form) database management and application development environment. Older implementation of the PICK environment were done with a proprietary o/s kernel called the PICK monitor. However, in last 7 years the PICK environment has become available with UNIX. The leading PICK/UNIX product today is one called uniVerse(tm) which is sold by Vmark Software. While there are some other PICK/UNIX offerings available, none to date has achieved the market acceptance of uniVerse. UniVerse is available on almost every major UNIX platform. Vendors represented include: IBM, Dec, AT&T, NCR, Pyramid, Encore, Sequent and many more. >2. Would it complement Ingres, for example, or completely replace it? It would completely replace Ingres. There is little 'productivity' software written under Pick, but there is a wide range of 'vertical' software in the marketplace (which nicely complements the productivity s/w which is now available under Unix). Some versions of Pick allow application development in several languages; all have BASIC interfaces. >3. Would Pick offer any advantages over major RDBMS's such as Ingres or Sybase? Depends upon your environment. PICK programmers aren't as plentiful as SQL types. OTOH it is much easier for novices to learn quickly and painlessly. Unless you're commited to what PICK can do for you it may not be wise to buck the system. There are two sites here at OSU that use PICK; it is nearly impossible to find PICK programmers here. We both end up training PICK programmers from scratch. Your area of the country may have more interest in PICK, and may have a better support base. It has been easier to program turnkey applications due to the strength of its VERY extended Basic language. The query language is easier to use than SQL. Native Pick does not do windows, communicate or do non-database things very well but a UNIX combination remedies these difficiencies. In general I've found PICK an excellent development environment. The data dictionary makes it easy to access the data, and the variable length fields are easy on our disk drive budget. Compared to Oracle, which we also have on our system, it is *VERY* easy to administer; I basically create an application and let our users go at it. PICK's BASIC programming language is more than just "BASIC". It provides excellent text handling features. BASIC has traditional arrays (x,y), Dynamic arrarys <f,v,s>, and text. Data types don't have to be declared; some cringe at this, and good programming practices are heartily encouraged to prevent problems. Conversion to and from numbers is automatic, but can be slow; tricks like "pie = 3.14159 + 0" forces a conversion to a number and the value retains that status. The language has FOR/NEXT, WHILE, LOOP, and CASE constructs. For FORTRAN fanatics it also has GOTOs, and Computed GOTOS. It also has a GOSUB construct which essential aids in building re-usable program modules. PICK doesn't use the SQL language to access data. It has its own command syntax (although the PICK standard is reportedly moving closer to SQL standards). A command like "SELECT filename WITH GENDER = 'F'" would be used to select specific data, or "LIST filename BY MAJOR BY GENDER BY NAME NAME FMT '30T'" would create a report, and with the "FMT" command, override the dictionary specifications. I like PICK a lot. It is flexible enough to do everything I've ever asked it to do. If you're willing to make a commitment to PICK it will work for you, but look carefully at your choices. [from Vmark Software:] 3. Frankly, those of us in the PICK industry feel that if PICK had been properly marketed in the past, there might not be products like Oracle, Ingres, Sybase or DB2 today. This is for three reasons: a. The PICK NFNF model allows a much more natural representation of real world complex data structures. Thus, programming applications is much easier since the data paradigm is easier to understand. b. PICK is an inherently simple system to use. Historically, PICK was by database users for database users. As a result, it contains many features that make it a joy to use. The query language of the PICK world is a simple dictionary driver interpreter that allows users to construct adhoc query in a natural way (i.e. LIST EMPLOYEES WITH SALARY > 50000 or LIST CLIENTS WITH NAME PRONOUNCED "SMITH" would produce the intuitively obvious reports). This ease of building adhoc reports has always been one the strong selling points of PICK. c. Although we have no hard data to date, there is strong anecdotal and empirical evidence to show that PICK has a definite performance edge over other environments. For example, I know of one sight in the UK that had originally purchased an HP system running uniVerse with plans migrate to Oracle. They have since decided to continue using uniVerse and not migrate to Oracle. They have cited two reasons for this decision: the uniVerse system is the most stable environment they have ever used, and to add 10 Oracle users to their system, they were told by Oracle that they would need to upgrade to a machine with 4 times the capacity -- the machine already supports 64 uniVerse uses. I hope these answers are of help and not too rambling in nature. If you would like some more information on PICK in general, or uniVerse in particular, please feel free to give me call. === James W. Todhunter [uvmark!jim@uunet.uu.net (Jim Todhunter)] === (Director, R&D)(Vmark Software, Inc.)(508-655-3700) [other contributors] Jerry M. Carlin [jmcarli@srv.pacbell.com] Cheers --dawn [ledbette@watserv1.uwaterloo.ca] Jeffrey R. Price [Hope this helps a little.][jprice@Prime.CoB.Ohio-State.edu] -- -ken rich ^()^ kenr@cc.rochester.edu