Contents of the APL-COMP.TXT file
A Shopper's Guide to Free and Low-Cost APL Interpreters/Demos
This article is a "Consumer's Report" on free or low-cost APL demos and
interpreters, with some pointers toward commercial APLs. It's a brief
scan, and many details are missing.
To begin, here are some speed and workspace size comparisons,
referenced to APL*PLUS, a widely used commercial APL interpreter:
Speed and Size: For several simple benchmark tests.
Speed (time in seconds, # is fastest) Size
Interpreter B1 B2 B3 B4 B5 CLEAR WS
APL*PLUS v10 5.1# 0.44# 0.66 22.2 0.29 341k
APL-I386 16.6 0.93 0.16# 21.6 0.02 1044k *
TryAPL2 21.5 1.09 2.75 7.9# 0.01# 233k
APL-IPC 32.4 1.06 0.44 31.1 0.07 195k
Sharp APL/PC " " " " " "
I-APL 74.9 13.62 17.49 226.8 25.59 32k
Benchmark B1: (2 + iota 10)*5: 10 exponentiations, looped 1000 times.
Benchmark B2: (iota 3000) divided by 44.3: 3000 f.p. divisions.
Benchmark B3: +/iota 15000: 15000 integer additions.
Benchmark B4: dyadic !: 20!40 looped 1000 times.
Benchmark B5: 40,000 boolean "ands" and "ors", not looped.
All times in seconds on a 386 33mHz machine, 4Mb RAM, no 387, DOS 5.0.
Times are the average of at least 3 runs. Sharp APL/PC, a shareware
interpreter, is similar to APL-IPC but without enhancements of APL-IPC
(editor, graphics, etc.). Sharp APL/PC is superceded by Iverson's APL-
IPC and APL-I386, and won't be discussed further.
* Default for this test machine. QuadWA for APL-I386 can be as large
as your RAM permits. QuadWA may be increased in APL*PLUS by a "virtual
Workspace" (disk-swapped). QuadWA for others is slightly configuration
dependent, except for I-APL, which is limited to 32k on DOS machines.
The range of speed and size is enormous: APL-I386 has a WS at least 32
times as large as I-APL (on test machine), and does integer addition 17
times faster than TryAPL2. Iverson APL-IPC and APL-I386 are both
faster than APL*PLUS for +/iota 15000. APL-I386 is 100 times faster
than I-APL for +/iota 15000. For Boolean operations, APL-I386 and
TryAPL2 are 15-30 times faster than APL*PLUS. For dyadic !, TryAPL2 is
3 times faster than APL*PLUS, and 28 times faster than I-APL. TryAPL
is an astounding 2500 times faster than I-APL for boolean "and".
These tests are not fully comparable: Iverson's APL-I386 uses the 32
bit data path of the 386 chip, whereas all others use only the 16 bit
data path. Even so, TryAPL2 is faster than APL-I386 on benchmarks B4
Only APL-I386 requires a 386 machine. All others work on both 386 and
lesser machines, but without taking advantage of 386 speed.
Among free/low-cost/demos, only IBM's TryAPL2 require a color monitor.
Can I do "real" work in any of these free/low-cost APLs?
APL*PLUS demo: No -- No access to user-defined functions/system commands
TryAPL2 : Yes -- with a few limits
APL-I386 : Yes -- 386 machine required
APL-IPC : Yes -- 386 machine not required
I-APL : No -- small WS, non-standard keyboard, slow
Price and Upgrade Path(s):
Disk $ Full $ 386
Interpreter Type price Documentation Product Version
APL*PLUS demo flat free* free/on-screen $695 $1700
APL-I386 nested $30 $30 - 3 books Is full product
TryAPL2 nested free* free + LJ print $500 free
APL-IPC nested $30 $30 - 3 books Is full product
I-APL flat free* $25 - 4 books Is full product
* Some suppliers charge a small fee. All marked * may be downloaded
free from the BBS\APL, 301-384-3672, 300/1200/2400b, N-8-1, 24 hrs.
Non-subscribers will need to ask the Sysop for permission to
IBM, Iverson APL-I386 and APL-IPC (and Sharp), and APL*PLUS II (386
version) all provide nested arrays. All are somewhat different in
their nested array features.
Keyboards for TryAPL2, APL2/PC (the full version of Try), and APL*PLUS
are all similar. These are the traditional APL keyboards. Iverson's
APL-IPC and APL-I386 keyboards are slightly different from IBM (and
from each other), but both resemble the APL*PLUS "unified" keyboard, as
used in the APL*PLUS demo. Most APL keys are in the same place,
differing mainly in the use of lower case, and in the use of ALT rather
than SHIFT to get special APL characters. Users of any of these APLs
can probably move from one keyboard to another easily. TryAPL2, the
APL*PLUS demo, and Iverson APL-I386 and APL-IPC have on-screen keyboard
I-APL is a different story. The I-APL keyboard is unrelated to any
other APL keyboard. APL learners who begin with I-APL must learn a new
keyboard when they upgrade to a more powerful APL. This is a drawback
for I-APL, although some attempts have been made to remap the keyboard
using public domain software.
A full screen editor is essential for modern APL programming. Among
commercial products, APL*PLUS and APL*PLUS II have a built-in assembly
language editor. Iverson APL-I386 and APL-IPC, and IBM APL2/PC, have a
full screen editor function that may be copied into your WS. TryAPL2 and
I-APL have only the outdated "del" editor.
Files: DOS Files APL Component files
APL*PLUS demo -- no no (all quad functions are inaccessible)
TryAPL2 -- no no (everything goes to AP211 files)
APL-I386 -- yes yes
APL-IPC -- yes yes
I-APL -- yes no
APL*PLUS demo -- On-screen keyboard tutorial/description of features
TryAPL2 -- On-screen documentation/lessons, LJ printable
APL-I386 -- Some on-screen, plus three books
APL-IPC -- Same as APL-I386 above
I-APL -- Very good on paper, essentially nothing on-screen
APL*PLUS demo -- Excellent
TryAPL2 -- Some, limited
IBM APL2/PC -- Some
APL-I386 -- Fairly complete
APL-IPC -- Same as APLI386 above
I-APL -- No, but users may write help files
Function Keys: settable on all except I-APL. Varying degrees of ease
of use on all others.
Non-English Language Keyboard/Character Support:
This feature is important only if you care about being competitive in
today's global marketplace.
APL*PLUS -- Very good for screen, nothing for keyboard
APL*PLUS demo -- N/A
TryAPL2 -- Excellent
IBM APL2 -- Excellent
APL-I386 -- Unknown (what works in APL-IPC doesn't work here)
APL-IPC -- Screen and keyboard, with some difficulty
I-APL -- No
Calls to Machine Language/C/Fortran etc.
APL*PLUS -- Yes
APL*PLUS demo -- No
TryAPL2 -- Unknown (probably yes?)
IBM APL2 -- Yes?
APL-I386 -- Yes?
APL-IPC -- no?
I-APL -- limited, machine language only
Vendor Support/Durability: Kaizen Strategy:* Responsive?
APL*PLUS -- Hot-line & bbs ** yes, to a fault highly so
TryAPL2 -- Hot-line, free yes apparently
IBM APL2 -- Hot-line, free uncertain unclear
APL-I386 -- Hot-line, toll yes apparently
APL-IPC -- Hot-line, toll yes apparently
I-APL -- none no no
* Kaizen = corporate commitment to unceasing, market-driven,
incremental improvement and innovation.
** Hot line is free for first 90 days, with varying levels of support
for varying prices thereafter.
IBM APL2 includes a free run-time system.
APL*PLUS and APL*PLUS II provide run-time systems at varying prices:
APL*PLUS 5-pack is $250, unlimited is $995; APL*PLUS II 5-pack is $875,
unlimited is $5000 -- plus $5 for shipping and handling.
Pricing strategies for Iverson Software (APL-I386, APL-IPC), and for
I-APL make run-time systems unnecessary.
Programming APL for run-time systems requires high skills in error
trapping. All APLs here, except I-APL, have error trapping.
Iverson APL-IPC and APL-I386 have built-in complex numbers. IBM's 386
version that comes with APL2/PC also has complex numbers.
Other: Graphics Color Printer Support
APL*PLUS and demo yes yes excellent
TryAPL2 yes yes v. good
APL2/PC yes yes unknown
APL-I386 yes yes v. good (LJ, PS, DM)
APL-IPC yes yes same as APL-I386
I-APL yes some limited
Iverson Software: APLIPC and APLI386:
Iverson Software Inc.
33 Major Street, Suite 466
Canada M5S 2K9 416-925-6096, Fax 416-488-7559
IBM : TryAPL2: (specify disk size)
IBM APL Development
M46/D12-278B, Santa Teresa Lab
Box 49023, San Jose CA 95161-9023
APL*PLUS demo : Manugistics, Suite 1729
2115 East Jefferson St.
Rockville MD 20852 301-984-5412
I-APL: U.S UK inquiries UK Orders
6611 Linville Dr. 2 Blenheim Rd. 59 The Crescent
Weed CA 96904 St. Albans Milton, Weston-super-mare
Herts AL1 4NR Avon BS22 8DU
This comparison of APLs is a collaborative effort among Dick Holt
(Silver Spring MD), Frank Ditto (Charlestown WV), Eric Iverson
(Toronto), and Les Moskowitz (Owings Mills MD).
Readers, please correct errors and omissions.