SHARP APL: Difference between revisions
Jump to navigation
Jump to search
Miraheze>Marshall No edit summary |
Miraheze>Marshall No edit summary |
||
Line 3: | Line 3: | ||
| index origin = 0 or 1 | | index origin = 0 or 1 | ||
| function styles = [[defined functions]], some [[tacit]] capabilities | | function styles = [[defined functions]], some [[tacit]] capabilities | ||
| numeric types = 64-bit float | | numeric types = 64-bit float, complex | ||
| unicode support = none | | unicode support = none | ||
| released = around 1979 | | released = around 1979 | ||
Line 11: | Line 11: | ||
}} | }} | ||
'''SHARP APL''' (later '''SAX''') was a standalone version of APL offered by [[I.P. Sharp Associates]] (IPSA), who had previously offered APL interpretation as a timesharing service. IPSA employed many notable APL designers including [[Ken Iverson]], and SHARP APL was the source of many developments in [[flat array theory]]. Notable features of SHARP APL include [[function rank]], the [[Rank operator]], and [[leading axis theory]], as well as [[close composition]] operators including [[Under]]. | '''SHARP APL''' (later '''SAX''') was a standalone version of APL offered by [[I.P. Sharp Associates]] (IPSA), who had previously offered APL interpretation as a timesharing service. IPSA employed many notable APL designers including [[Ken Iverson]], and SHARP APL was the source of many developments in [[flat array theory]]. Notable features of SHARP APL include [[function rank]], the [[Rank operator]], and [[leading axis theory]], as well as relative [[comparison tolerance]] and [[close composition]] operators including [[Under]]. | ||
== Primitive functions == | == Primitive functions == | ||
Line 215: | Line 215: | ||
=== Numeric types === | === Numeric types === | ||
SHARP | SHARP originally supported only real numbers using double (8-byte) precision. Numbers were stored in one of three types: | ||
* Boolean, with one bit per value | * Boolean, with one bit per value | ||
* Integer, with four bytes per value | * Integer, with four bytes per value | ||
* Floating, with eight bytes per value | * Floating, with eight bytes per value | ||
[https://www.jsoftware.com/papers/satn40.htm SATN-40] describes the addition of complex numbers to SHARP APL. | |||
== External links == | == External links == | ||
[https://en.wikipedia.org/wiki/I._P._Sharp_Associates Wikipedia] | [https://en.wikipedia.org/wiki/I._P._Sharp_Associates Wikipedia] |
Revision as of 13:16, 31 October 2019
SHARP APL (later SAX) was a standalone version of APL offered by I.P. Sharp Associates (IPSA), who had previously offered APL interpretation as a timesharing service. IPSA employed many notable APL designers including Ken Iverson, and SHARP APL was the source of many developments in flat array theory. Notable features of SHARP APL include function rank, the Rank operator, and leading axis theory, as well as relative comparison tolerance and close composition operators including Under.
Primitive functions
Arithmetic
Scalar
All scalar functions have rank zero.
Glyph | Monadic | Dyadic |
---|---|---|
+ |
Conjugate/Identity | Addition |
- |
Negate | Subtraction |
× |
Signum | Multiplication |
÷ |
Reciprocal | Division |
* |
Exponential | Power |
⍟ |
Natural Logarithm | Base-⍺ Logarithm |
| |
Magnitude | Residue |
! |
Factorial | Out-Of/Combinations |
⌊ |
Floor | Minimum |
⌈ |
Ceiling | Maximum |
○ |
Pi Times | Circle functions |
∧ |
And/Least Common Multiple (LCM) | |
∨ |
Or/Greatest Common Divisor (GCD) | |
⍲ |
Nand | |
⍱ |
Nor | |
~ |
Not | See Miscellaneous |
? |
Roll | See Non-scalar |
Non-scalar
Glyph | Rank | Monadic | Ranks | Dyadic |
---|---|---|---|---|
⌹ |
2 | Matrix inverse | ∞ 2 | Matrix divide |
? |
See Scalar | * * | Deal | |
⊤ |
∞ ∞ | Encode | ||
⊥ |
∞ ∞ | Decode |
Relational
Glyph | Rank | Monadic | Ranks | Dyadic |
---|---|---|---|---|
= |
See Miscellaneous | 0 0 | Equals | |
≠ |
See Miscellaneous | 0 0 | Not Equals | |
< |
See Structural | 0 0 | Less Than | |
≤ |
0 0 | Less Than or Equal | ||
≥ |
0 0 | Greater Than or Equal | ||
> |
See Structural | 0 0 | Greater Than | |
≡ |
∞ ∞ | Match | ||
∊ |
0 ∞ | Membership | ||
⍷ |
∞ ∞ | Find |
Indexing
Glyph | Rank | Monadic | Ranks | Dyadic |
---|---|---|---|---|
@ |
See Miscellaneous | 0 ∞ | From | |
⍳ |
1 | Count | 1 0 | Index Of |
⍸ |
∞ ∞ | Index | ||
⍋ |
∞ | Numeric Grade up | ∞ ∞ | Character Grade up |
⍒ |
∞ | Numeric Grade down | ∞ ∞ | Character Grade down |
Structural
Glyph | Rank | Monadic | Ranks | Dyadic |
---|---|---|---|---|
⍴ |
∞ | Shape of | 1 ∞ | Reshape |
↑ |
See Miscellaneous | 1 ∞ | Take | |
↓ |
∞ | Raze | 1 ∞ | Drop |
< |
∞ | Enclose/Box | See Relational | |
⊃ |
∞ | Conditional Enclose | ∞ ∞ | Link |
> |
0 | Disclose/Open | See Relational | |
, |
∞ | Ravel | ∞ ∞ | Catenate |
⍪ |
∞ | Table | ∞ ∞ | Catenate-Down |
⌽ |
1 | Reverse | 0 1 | Rotate |
⊖ |
∞ | Reverse-Down | ∞ ∞ | Rotate-Down |
⍉ |
∞ | Monadic Transpose | 0 ∞ | Dyadic Transpose |
Miscellaneous
Glyph | Rank | Monadic | Ranks | Dyadic |
---|---|---|---|---|
⊣ |
∞ | Stop | ∞ ∞ | Left |
⊢ |
∞ | Pass | ∞ ∞ | Right |
≠ |
∞ | Nubsieve | See Relational | |
↑ |
∞ | Nub | See Structural | |
= |
∞ | Nubin | See Relational | |
~ |
See Scalar | ∞ ∞ | Less | |
@ |
1 | All | See Indexing | |
⍕ |
∞ | Monadic Format | * ∞ | Dyadic Format |
⍎ |
* | Execute |
Primitive Operators
Glyph | Operands | Ranks | Monadic Call | Dyadic Call | ||
---|---|---|---|---|---|---|
/ |
f |
∞ |
Reduce | |||
⌿ |
f |
∞ |
Reduce-down | |||
\ |
f |
∞ |
Scan | |||
⍀ |
f |
∞ |
Scan-down | |||
/ |
m |
∞ |
Copy/Compress | |||
⌿ |
m |
∞ |
Copy-down/Compress-down | |||
\ |
m |
∞ |
Expand | |||
⍀ |
m |
∞ |
Expand-down | |||
⊂ |
f |
∞ |
rf lf |
Swap | ||
& |
f |
∞ |
∞ ∞ |
Select | ||
⍤ |
f |
g |
mg |
mg mg |
On (close Over) | |
⍤ |
f |
n |
n |
n n |
Rank | |
⍤ |
m |
g |
mg |
mg mg |
Cut | |
⍥ |
f |
g |
mg |
mg mg |
Upon (close Atop) | |
¨ |
f |
g |
mg |
mg mg |
Under | |
¨ |
m |
g |
mg |
With | ||
¨ |
f |
n |
mf |
With | ||
. |
f |
g |
2 |
∞ ∞ |
Alternant | Inner-product |
. |
m |
g |
∞ ∞ |
Tie, Outer-product | ||
. |
f |
m |
mf |
Ply |
Implementation
Numeric types
SHARP originally supported only real numbers using double (8-byte) precision. Numbers were stored in one of three types:
- Boolean, with one bit per value
- Integer, with four bytes per value
- Floating, with eight bytes per value
SATN-40 describes the addition of complex numbers to SHARP APL.