OpenVMS Demos: Programmer's Corner (old)

OpenVMS Resources (non-code) can be found here

OpenVMS Application Demos

OpenVMS: Source Code in HTML
Click here to view a directory of my zipped files not yet converted

"HP-BASIC for OpenVMS Alpha"

(a.k.a. "Compaq-BASIC for OpenVMS Alpha", "DEC-BASIC")
(should work with no modifications on VAX or Itanium)

<form method="post" enctype="multipart/form-data" action="/scripts/upload_test_neil">
  <input type="text"   name="textline">
  <input type="file"   name="datafile">
  <input type="submit" name="Send">
</form>

System Password Notes:

1. OpenVMS really is "open" because I wrote the afore mentioned program in a about an hour using two OpenVMS system calls (sys$getjpi and sys$hash_password) which are publicly documented here. You need to have system privileges in order to access someone else's hashed password, but nothing could prevent an unscrupulous system manager (or support staff) from using this technique to acquire someone else's password in order to commit a crime using that person's account. This is a big no-no in the world of Sarbanes-Oxley (SOX)
    
2. My busy little AlphaServer-4100 (EV5/300MHz) averages 1 million guesses in 273 seconds. Since there are 38^6 "six character passwords", a worst case run would require 3,010,936,384 guesses. Divide this number by 1M and you get 3010.9 which then is multiplied by 273 then divided by 3600 to yield 228 hours or 9.5 days. Multiply this figure by 38 for every character added to the six character password, so a seven character password would require 361 days while an eight character password would require 37.5 years thus forcing criminals to look for another method.
    
3. as a result of this experiment, I've decided to set the minimum password length to 8 on all of my OpenVMS systems via the following command in authorize:

    $set def sys$system
    $run authorize
    uaf>mod * /pwdminimum=8
    exit
    $

   which results in the total number of possible candidate minimum length passwords to be guessed changing from 38^6 (3.01e+9) to 38^8 (4.34e+12). Of course you could always use a larger password and I think I'll do this on my priv accounts. BTW, it might be a good idea to monitor unusual access to SYSUAF using ACL's and $SET AUDIT
    

4. currently any password you enter is UPCASED before being tested and/or stored. Starting with OpenVMS-7.3-2 an authorize flag named "PWDMIX" to enable case-sensitive and extended-character passwords.
    
5. password cracking can be sped up considerably by copying the SYSUAF to a PC (you still need OpenVMS system privs to do this) and running a standalone tool. Check out this link for more details (p.s. tools like this are one more reason to enforce longer passwords and enable the "case-sensitive password feature" which will be available with  OpenVMS-7.3-2)

HP-BASIC for OpenVMS VAX

(a.k.a. "Compaq-BASIC for OpenVMS VAX", "Compaq-BASIC for VAX", "VAX-BASIC")

• download a demo which calls a MACRO-32 function to PEEK at memory (can be converted to POKE but I can't think why you'd want to do this)

HP-CXX for OpenVMS Alpha
HP-CXX for OpenVMS Itanium
HP-CXX for OpenVMS VAX

(a.k.a. "Compaq-CXX for OpenVMS Alpha", "Compaq-C++", "HP-C++")

• download a C++ application to convert SEGY formatted data to ASCII (then you will be able to pull the data into a spread sheet). Note that SEGY is an acronym for "Society of Exploration Geophysicists" Y-format. This name of this program is SEGY2ASC

HP-C for OpenVMS Alpha
HP-C for OpenVMS Itanium
HP-C for OpenVMS VAX

(a.k.a. "Compaq-C for OpenVMS Alpha", "Compaq-C for OpenVMS VAX", "DEC-C")

Note: VAX-C was phased out prior to VMS 6.0 and replaced with DEC-C. Don't do any new program development with VAX-C.

DCL (Digital Control Language a.k.a. DEC Control Language)

M32 (a.k.a. "Macro for OpenVMS VAX", "Macro-32")

• download a little utility to reset hardware error counters in OpenVMS (unless we reboot, this is all we have until OpenVMS-7.3-2)

M64 (a.k.a. "Macro for OpenVMS Alpha", "Macro for OpenVMS Itanium", "Macro-64")

• learn about HP's MACRO64 freeware (for Alpha) including the required free license to activate it.
  Note: OpenVMS Engineering people have told me that HP-C has replaced MACRO-64 as the preferred way to generate Alpha binaries.

My 2 cents worth about "HP-BASIC for OpenVMS Alpha"

1. Native (built-in) support for RMS-based files including sequential, relative and indexed.
   Yes, you can develop non-SQL compliant databases.
    
2. Like Pascal, run-time errors can (if desired) be handled locally with WHEN ERROR blocks rather than using a common handler like "ON ERROR GOTO". But you can use both methods if you desire.
    
3. Enforced variable declaration can be achieved with the directive "OPTION TYPE=EXPLICIT"
    
4. SELECT CASE statements seem much more readable than compound "IF THEN" - "ELSE" - "END IF"
    
5. Line numbers are not required (unless you reference the BASIC function "ELN")
    
6. Interpreter mode is only available on VAX
    
   
7. When Microsoft shifted from VB6 to VB.net, their implementation of BASIC became more object-oriented. VMS-BASIC is still very procedural.

The Name Game

• In the 1990s, DEC (Digital Equipment Corporation) introduced the 64-bit Alpha platform meant to eventually supplant 32-bit VAX.
  • So VAX-BASIC was rebranded DEC-BASIC for VAX
  • A new product called DEC-BASIC for Alpha was released
• In 1998, Compaq acquired DEC so this two-platform language was rebranded Compaq-BASIC for VAX and Compaq-BASIC for Alpha.
• In 2002, HP acquired Compaq so the products were renamed HP-BASIC for VAX and HP-BASIC for Alpha (perhaps the name should be changed to VMS-BASIC)
• In 2004, HP started shipping Itanium systems (meant to eventually supplant Alpha). Now we've got a third product called HP-BASIC for Itanium

1. Compile mode only; no interpreter (I do not consider this a problem)
    
2. The Alpha compiler does much better code optimization than VAX so if you weren't really careful while creating code for VAX just recompiling for Alpha will provide you with compile-time speed improvements. For example, the following two programs (example 2a and 2b) generate different object code on VAX but generate the same optimized object code on Alpha even when complied with optimization=level=0 (none).

                             ! example 2a (written by a careless VAX programmer).
   1000 a% = 3               ! the VAX complier produces CISC code to convert 3 from real to integer at runtime
        b% = 4               ! the VAX complier produces CISC code to convert 4 from real to integer at runtime
        print a% + b%        !
        end program 1        ! return -s- to DCL
   
                             ! example 2b (written by a careful VAX programmer).
   1000 a% = 3%              ! both compilers generate code which represents an integer 3
        b% = 4%              ! both compilers generate code which represents an integer 4
        print a% + b%        !
        end program 1        ! return -s- to DCL

3. However, running the Alpha compiler with optimization=level=4 (full) would yield the following equivalent program:

                             ! example 3 (why add the variables if you don't have to?)
   1000 print "7"            ! the compiler added 3 to 4 "at compile time" so only this I/O statement was required
        end program 1        ! return -s- to DCL

4. Here's another example program which yields radically different object files on VAX compared to Alpha.

   1000 option type=explicit ! example 4
        declare long i%      !
        for i% = 1 to 100    ! should be "1% to 100%" on VAX; Alpha doesn't care because long was declared
            print i%         !
        next i%              !
        end program 1        !

   If you compile with optimization=level=4 and used a smaller loop, the Alpha complier might decide to just generate a series of PRINT statements. This is called "loop unrolling".

5. So-called "DEC languages" targeted at Alpha and Itanium are based upon the "GEM Optimizing Compiler System". GEM is a project code-name rather than an acronym. Everything coming out of DEC in those days had names like GEM, Emerald, Opal, etc.
   A Few GEM links:
   1. http://www.hpl.hp.com/hpjournal/dtj/vol4num4/vol4num4art8.pdf
   2. http://source.rfc822.org/pub/mirror/ftp3.ds.pg.gda.pl/people/macro/DEC/DTJ/DTJ808/DTJ808PF.PDF
   3. http://h71000.www7.hp.com/openvms/journal/v5/portingmacrocompiler.html

Other Noteworthy Applications

Intercepting OpenVMS Mail Delivery (using DELIVER)

• I'm amazed by the number of OpenVMS people who are not aware of the DELIVER freeware program which allows you to intercept "OpenVMS Mail Delivery" (including mail from TCP/IP sources) and then do cool things like run user-written programs and/or batch jobs to automatically process the message.
• The distribution includes Pascal source code and binaries for both VAX and Alpha.
• It can be found on the V4 Freeware CD-ROM shipped with OpenVMS or can be downloaded free of charge here at the HP website or here at Process Software's OpenVMS resource area
• OpenVMS Notes: DELIVER (the email interception utility)

Intercepting OpenVMS Mail Delivery (using PMDF)

• If you don't want to use unsupported freeware to intercept your email, then checkout PMDF from Process Software which can also do many other cool things like MIME DECODE etc.
• Binaries are also available for UNIX and Windows-NT (including Windows-2000). I can only assume that Process Software will be producing an Itanium version after they port TCPware.
• if you trial this software, be sure you execute DCL command $PMDF CONFIG before you begin testing

External Links:

• my OpenVMS resource page
• my OpenVMS Mini-FAQ page
• my "VAX to Alpha" Porting Diary page

Back to Home
Neil Rieck
Kitchener - Waterloo - Cambridge, Ontario, Canada.