Still using my 30 yr old Northgate OmniKey Ultra every day. I'd buy another in a hearbeat if I could.

Albert-
I'm confused as to what you are doing in this thread.
Are you saying you took the above code and compressed a file by 92%, then decompressed the result using the above code, and the input and output files were identical?

It's a huge pig at 2.5 megs, and it probably uses lots and lots of different api's to get the job done. If you decide to use it, use the standalone portable version, not the installable version. There is no reason to install it at all. Just unzip it to a folder and run it manually or drop a shortcut...

I use ShellFolderFixUI as a startup program to remember how I want each folder to show. Works great.

albert- The data we have been using for testing, random numbers of the range 0-255, is not compressible, otherwise it wouldn't be very good random numbers. For example, I ran the generator for str1 and saved it to a file. That file was 190,000,000 bytes long. I then ran that file through 7-zip to co...

I lifted your string generator to be sure we were working on the same input. I ran your last assembly version which I called tally, and my last version that wasn't reading a file (instrtest11) for comparison. This is what I got when running in a command window from a dos prompt: F:\FreeBasicProgs\Te...

Thank you. It never occurred to me that an array would be on the stack.

This program just dies after the first input without going to the error routine, so I can't figure out what the error is. I'm sure it's obvious, I just can't see it. Any help will be appreciated. ' compile with -exx On Error GoTo fatal Dim erf As Long Dim As ULong dictcount(255,255) Dim As String di...

albert- I know I said I was done, but I couldn't resist. Here is a program that reads in a file and creates the dictionaries. Using a 13 meg test file, on my machine ( with plenty of memory and fast cpu) it takes a grand total of 0.3 seconds. Is that fast enough? ( it did take 260 seconds for a 1.1 ...

And finally, albert, this one uses separate dictionaries for each of 256 possible starting letters (dicts(255)). It runs in 8 seconds, so it takes longer to compile than to run. I'm done here unless you have other questions. '#Define debug Dim As ULong a,i,b,j,testsize Dim As UByte uu,vv,yy,onetime ...

That last one takes 779 seconds on my computer. By sorting the dictionary so words starting with the same letter are grouped together, and only searching within the group needed, it take 16 seconds, almost down to usable speed. Here's the test program Dim As ULong a,i,b,j Dim As UByte uu,vv,yy,oneti...

Using integers would look something like this. Still not terribly fast, but I don't see how you could run any faster using brute force code. To get any more speed, you would have to do some kind of hashing. Perhaps make 256 dictionaries, where all words in a dictionary would start with the same lett...

If I understood you correctly, then your instr is not the right thing to use. As I showed in my example, it will find a match at byte two, which is not on a 4-byte boundary. If everything is to be in units of four bytes, then it should be faster to do everything in four-byte integers (longs), rather...

albert- with your instr instruction, you are testing for any occurrence of a four byte string. Did you mean for the match to occur anywhere within the string, or only at four byte boundaries? e.g. if str1 looks like- abcdefghijklmnopbcdefghi 1 2 3 4 5 6 7 8 9 and you pick up the fifth group (bcde) a...

a$[c,d] Won't work because: a) string index operator [] can only take one argument (index). b) this means, you'll need another construct (given by the language) Precisely what I'm proposing. A compiler update to allow the syntax of [startcol, optional end col] That's all I ever asked for. c) conclu...