7ZIP FORMAT.... (possible implementation in TC??)

[SanskritFritz]

I remember UC2 and, back in the old days, it used to be my primary compression program. I am sad that development has stopped because it was quite good, reliable, and had features then that other compressors didn't have (but do now).

UC2 can hold more versions of a file. Which packer can do this today? (Keep-it can, but it doesnt work on NT based systems )
And speaking about 7zip, it seems to be our new toy

[Clo]

Randy

 Hi again !

…a 1.4ghz machine with 256m of RAM, 7-Zip quickly climbed to using around 300 megs of RAM and nearly every CPU cycle. …

• It's also dependent of the types and sizes of the files to pack.
* To get a smaller output (packed) file, the largest lexicon as possible must be built and used. That needs complex calculations, hence a high consumption of memory and CPU-use. Such lexicons having large bases for the longest redundant strings as possible are more greedy for all, indeed.
* For instance, whether you have big text-files to pack, let's say : some in French, others in English, and others in the Coco-Islands language , the packer must build three different lexicons, and use them as and when needed. The same for binary-files when you have several kinds of them.

…I haven't tried it on my workstation (3ghz, 1 gig RAM) but it seemed excessive. …

• Yes, it's excessive. But I'm pretty sure that Ch. Ghisler is quite able to optimize the running of 7-ZIP if used into TC as minimal pack / unpack DLLs or so… Since the interface of TC should be used directly, that'ld save a lot of resources.
* I've a workstation so powerful too : 2-CPU @ 1.65 GHz - 1 GB DD-RAM - XP-Pro (mandatory to manage two CPUs).
* But I bet that such an additional packer could run properly in TC with my old P II @ 350 MHz - 320 MB of SD-RAM - Win 98 SE (Fr.)
- It's my usual PC, I much prefer to work with it, rather than XP and its useless thick layer of make-up
Well, "Wait & see…", but I'm curious to give a try to such an addon, maybe in the next major-version (6.5, I guess) ?

P.S. France rules!

Kind regards,
Claude
Clo

[norfie]

[norfie]

[Clo]

Randy & all…    

Hello !

• A test like the Norfie's one, but using 7-Zip with its own file manager interface, 'cause I didn't install MultiArc.
* With the P II @ 350 MHz quoted above.
• Packing a whole program-directory, Corel30 = 22.7 MB - All kinds of files, number = 558 files and / or sub-dirs. All 8.3 file-names…

1. Corel30.7z > 5' 5.8 MB 7-Zip 3.13
2. Corel30.ace 1' 32" 6.4 MB WinAce 2.1.1.0
3. Corel30.rar 1' 15" 7.5 MB WinRar 3.0
4. Corel30.uc2 ~ 2' 8.0 MB
5. Corel30.zip ~ 2' 8.5 MB Internal TC Zip 
6. Corel30.arj ~ 2' 8.5 MB ARJ-32
7. Corel30.lzh ~ 1' 50" 8.7 MB LHA 32

• About the output archive-size, there is not finish-photo !
…if you are not hurry too much !
* Even the dear old uc2 is better at size than the built-in TC-Zip…
* Currently, the best compromise time / output-size is WinAce in that case. But isn't free.
• I would like 7-ZIP in TC directly, when the smallest archive-size is the main criterion.

K R
Claude
Clo

[norfie]

[norfie]

[Maxwish]

What does CAB do in your tests ?
I find it surprisingly good sometimes

[Clo]

2Maxwish
Hello !
• Right, but TC doesn't manage *.CAB at packing. I guess it's a M$' patent issue (?).
• I can't pack directories as CAB in that PC.
- With a single file, I get:
* Logfile.txt = 108.9 KB
* Logfile.cab = 8.5 KB
* Logfile.zip = 8.4 KB  (TC internal ZIP)
• I used a small 3rd.-party program aCAB; I'll contact its author, in order he improves the program to pack whole directories (I hope...)

  KR
Claude
Clo

[norfie]

[lzvk25]

Why don't you take a look at :
http://www.maximumcompression.com/

There they have test for several compressors even experimental ones. I even played with some of them.

[norfie]

[Sheepdog]

I don't like the results.

Why, don't they fit your prejudice?

sheepdog

[icfu]

Windows XP has a built-in cab-packer:
iexpress.exe

Icfu

[Randy]

Everyone, thanks for all of your input. I can see that 7-Zip clearly has an edge over most compressors if time and CPU cycles aren't a concern. How does it fare when unpacking a tightly compressed file? I imagine that it wouldn't take nearly as long or be as CPU-intensive as creating the archive. If this is the case, I can see where 7-Zip would be excellent at creating long term archive solutions where maximum compression is necessary although I'm not sure I could see it being useful for day-to-day on-the-spot compressions (although I might change my mind if it were properly integrated into TC).