You can write an external memory spell checker with a tiny amount of RAM: something like
- sort the words in the document
- eliminate unique words (they sort together)
- merge the sorted words with the sorted dictionary and keep only the missing words
I saw this in BASIC in Creative Computing and got it working in on my TRS-80 Color Computer which had much less than 32k of available RAM, so that was the first thing I thought when I saw the headline.
it had a compressed dictionary that would fit together with the other programs you were running on a PC and spell check as you typed; there was a 640k limit for the PC but it could only use a fraction of that so as not to interfere and in the early days of the PC you couldn't actually afford to fill it out.
I can't remember the name of the product but in the 80s there was a hardware spell checker for the IBM PC. It was a box that connected between your keyboard and your PC, and if you ever typed a string of letters that it did not recognize as a dictionary word, it would beep to let you know.
One of the things that got me intrigued by Unix was an early 1980s(ish) Byte article which walked through building a (trivial example, not the "real" one) spell checker out of a split/sort/comm pipeline, something like 7 commands? 8-bit PCs didn't have anything like that, and yet it didn't look like it needed that much sophistication...
in the mid 80's i ran into something similar. Fast is relative.
I had a lot of data, 640KB RAM, 64KB of heap, and 64KB of stack. I had hundreds of megabytes that I had to search extract data from and then combine some of them.
I experimented with data index structured into ternary trees. Conceptually it made sense, but implementation-wise the relationships and paths were still too big to keep in 64KB.
Instead of compression, I did swapping. I wrote a TSR (think service), a piece of code that would process a chunk of the data, extract the results, store it n the stack, dump the original data, make an interrupt call to the TSR, which in turn destroy the heap, and read in the next chunk from storage, return control to the program, process, combine with stack data, and continue until finished the entire process.
Originally this process took about a week for three data entry persons (think about a dozen 3" ring binders filled with tables), and an specialist combining the information. The program completed the work in just a few hours. It was amazingly "fast".
For perspective, in 1983 or so, Grammatik on CP/M ran in under 64k and did "grammar checking" (spell checking, plus a bunch of expert system rules) on an 8-bit system. (It sticks in my memory because of the time spent poking at the really interesting part: that it was so compact because it was in Forth, and there was enough of an outer interpreter in the product that with a little hex editing you could just use it as a Forth interpreter - with a very specialized set of functions preloaded :-)
How about 39kB for a video game with physics, dynamic graphics, two music tracks, sound effects, online high scores, and built-in instructions? https://news.ycombinator.com/item?id=38372936
Sizecoding is a thing. .kkrieger is a rather famous 96kB FPS game. There is even an entire demoparty called Lovebyte that is dedicated to it, the biggest category is 1k, but all demoscene events I can think of have a sizecoding competition of some kind.
And it is a completely different thing. In general, it is more about procedural generation and tricks then good packing. Runtime packers are used, like crinkler and kkrunchy, but actually they use a lot of RAM, like hundreds of MB, which is a bit surprising considering that the decompressed executable is in the tens of kB. But that's because they use very powerful but slow compression algorithms.
Sizecoding usually doesn't care about RAM, unless the platform requires it, the only think that matters is the size of the executable file and its data. For that 39kB Playdate game, I guess that's the same idea. The Playdate has 16MB of RAM, I bet the game took full advantage of it.
Also for reference, for those who aren't familiar with typo and don't want to read the C source code listed above (side-comment: citing the comment-free source code 'for reference' is hilarious. Thank you for the laugh :) )
You can write an external memory spell checker with a tiny amount of RAM: something like
I saw this in BASIC in Creative Computing and got it working in on my TRS-80 Color Computer which had much less than 32k of available RAM, so that was the first thing I thought when I saw the headline.Now this blew people away when it came out
https://winworldpc.com/product/turbo-lightning/1x
it had a compressed dictionary that would fit together with the other programs you were running on a PC and spell check as you typed; there was a 640k limit for the PC but it could only use a fraction of that so as not to interfere and in the early days of the PC you couldn't actually afford to fill it out.
Link seems to be broken.
I can't remember the name of the product but in the 80s there was a hardware spell checker for the IBM PC. It was a box that connected between your keyboard and your PC, and if you ever typed a string of letters that it did not recognize as a dictionary word, it would beep to let you know.
One of the things that got me intrigued by Unix was an early 1980s(ish) Byte article which walked through building a (trivial example, not the "real" one) spell checker out of a split/sort/comm pipeline, something like 7 commands? 8-bit PCs didn't have anything like that, and yet it didn't look like it needed that much sophistication...
in the mid 80's i ran into something similar. Fast is relative.
I had a lot of data, 640KB RAM, 64KB of heap, and 64KB of stack. I had hundreds of megabytes that I had to search extract data from and then combine some of them.
I experimented with data index structured into ternary trees. Conceptually it made sense, but implementation-wise the relationships and paths were still too big to keep in 64KB.
Instead of compression, I did swapping. I wrote a TSR (think service), a piece of code that would process a chunk of the data, extract the results, store it n the stack, dump the original data, make an interrupt call to the TSR, which in turn destroy the heap, and read in the next chunk from storage, return control to the program, process, combine with stack data, and continue until finished the entire process.
Originally this process took about a week for three data entry persons (think about a dozen 3" ring binders filled with tables), and an specialist combining the information. The program completed the work in just a few hours. It was amazingly "fast".
This was on a single threaded system.
[0] https://en.wikipedia.org/wiki/Terminate-and-stay-resident_pr...
For perspective, in 1983 or so, Grammatik on CP/M ran in under 64k and did "grammar checking" (spell checking, plus a bunch of expert system rules) on an 8-bit system. (It sticks in my memory because of the time spent poking at the really interesting part: that it was so compact because it was in Forth, and there was enough of an outer interpreter in the product that with a little hex editing you could just use it as a Forth interpreter - with a very specialized set of functions preloaded :-)
Way too smart for worse is better. Think Worser!
The main memory bandwidth and the disk bandwidth were about the same, a little of 1MB/s.
I would have done this in multiple passes (but still used the Bloom Filters, those are cool).
https://github.com/arnoldrobbins/v10spell
https://code.google.com/archive/p/unix-spell/
The original paper is great https://www.semanticscholar.org/paper/Development-of-a-Spell...
It is hosted on his web page https://www.cs.dartmouth.edu/~doug/
https://en.wikipedia.org/wiki/Douglas_McIlroy
If you are a word nerd, you will have found obovate and there, this chart.
https://upload.wikimedia.org/wikipedia/commons/e/e8/Leaf_mor...
I wonder what common typos this misses thanks to the hashing?
Related, a contest about compressing the Wordle dictionary: http://golf.horse/wordle/
How about 39kB for a video game with physics, dynamic graphics, two music tracks, sound effects, online high scores, and built-in instructions? https://news.ycombinator.com/item?id=38372936
Sizecoding is a thing. .kkrieger is a rather famous 96kB FPS game. There is even an entire demoparty called Lovebyte that is dedicated to it, the biggest category is 1k, but all demoscene events I can think of have a sizecoding competition of some kind.
And it is a completely different thing. In general, it is more about procedural generation and tricks then good packing. Runtime packers are used, like crinkler and kkrunchy, but actually they use a lot of RAM, like hundreds of MB, which is a bit surprising considering that the decompressed executable is in the tens of kB. But that's because they use very powerful but slow compression algorithms.
Sizecoding usually doesn't care about RAM, unless the platform requires it, the only think that matters is the size of the executable file and its data. For that 39kB Playdate game, I guess that's the same idea. The Playdate has 16MB of RAM, I bet the game took full advantage of it.
Not the same when the machines in question are not on the same level...
This spell check must have come later. The original spell check burst the words of the document, and sort -u'd them:
makewords sentence | lowercase | sort | unique | mismatch
64kB was huge. The first version of UNIX needed 24kB (half of which was taken by the kernel).
Marginally related... has anyone ever ported the "typo" program to modern C?
For reference, https://github.com/robpike/typo/blob/master/unix/typo.c
Also for reference, for those who aren't familiar with typo and don't want to read the C source code listed above (side-comment: citing the comment-free source code 'for reference' is hilarious. Thank you for the laugh :) )
https://github.com/robpike/typo/blob/master/typo.png
I had spelling checkers on the Apple ][ that ran in 48K!
Exactly. 64K only seems tiny to people who didn't use home computers in the 1980s.
Cool. Now move forward.