Разнообразно приветствую тебя, Valentin!
12 Октября 2025, Valentin Kuznetsov писАл к Bender Rodriguez следующее:
окт 12 1992. Charles Buchanan (1:3812/10.6) пpовёл
тестиpование двенадцати pазличных тоссеpов и опубликовал
pезультаты. По скоpости pаботы пеpвое место занял Fastecho
1.20a.
http://st.g0x.ru/haroshaya_shtuka.png
Всмысле список и все тоссеpы из него Ж+)
FidoNews 9-41
12 Oct 1992
===
By Charles Buchanan 1:3812/10.6
A Comparison of Mail Tossers
I guess that I better start off by telling you what is meant by a
mail tosser. All of the echomail that goes throughout Fidonet is sent
along in mailbags from different systems with all of the messages from
all the different echomail areas bundled together. They are not
separated and sent each in their own area. So when a mailbag arrives,
it has one or several *.PKT files inside of it that need to be sorted
and put into the correct echomail area. The mail tosser is the
program that puts the messages into the correct areas as well as into
the correct directories on your disk. It will also create indices or
at least a supplemental utility will. This is a very basic
explanation of a mail tosser and what it does. Some mail tossers have
alot of different features but I don't plan to talk about that.
There are 3 different types of message bases -- Fidostyle, Hudson,
and Squish. Each of the three types has their own good points and bad
points. So I guess that is why there are so many different mail
tossers around. Here is general overview of what these 3 different
types of messagebases are :
Fidostyle -- Messages from each echo (or conference) are in separate
files as well as separate directories. Each of the messages are in
a separate file that are of the form *.MSG
Hudson -- All of the messages are kept in one file and the indices,
lastread pointers, and users are kept in separate files. So the
actual message base has all of the messages from all echos in one
file. The format of the Hudson message base is *.BBS for all of the
files needed.
Squish -- Squish is sort of a cross of the two types above. Squish
keeps each echo separate in it's own messagebase but they can all be
in the same directory. Squish also keeps separate dupes, and pointers
for each messagebase.
That is very basic explanation of the three types of messagebases
that I'm just barely familiar with and how I understand them. I'm an
expert by no means and I welcome clarification on anything that I've
gotten wrong.
I operate as a point and I use a mail tosser for all of the mail that
comes into me. After I was struck by the infamous Feb 29 bug that
made some mail tossers choke, I started looking around at what was
available. I noticed that there are quite a few mail tossers. Some
have more features than others. Some are very simple and others are
quite complicated.
One of my favorite echos that I like to participate in is POINTS.
There was a discussion a few weeks ago about different tossers,
speed, features, space used and other things. I was told that Squish
actually uses less diskspace for the message base than Hudson. So I
kept a couple of my old mailbags to try Squish for myself. Then
tossed the same mailbags with a Hudson style tosser. There was indeed
a difference in the speed and the space used by both. This then got
me curious about other mail tossers and how they compared ? I also
wondered if all Hudson style mail tossers were the same ? Well, I was
surprised by what I found out !!
My comparison in this article is really limited to three criteria --
speed, space used for the messagebase, and point awareness. The point
awareness is by neccessity since I operate as a point and all of my
old mailbags have my address in them as a true 4-D address. This
means that I could not try a tosser that either -- 1) Would not
support 4-D addressing or 2) Required a pointnet or a fakenet
addressing scheme. I also did not want to compare all of the bells
and whistles and how many of each they all had. When I looked at
these tossers, I set them up as I would use them -- maybe not how
other people would use them. So this means that this comparison is
not real scientific nor is it the most complete comparison. But speed
and diskspace usage are a common factor in all of them.
Now I'll explain my setup and how I did my comparisons. I setup each
of the mail tossers and made sure that they would work for me. Then I
created a batch file that would try each of them out and recorded the
results. Since the unarchiving of the mailbags would always be the
same, I did that first and just kept the *.PKT files in my inbound
directory. I also deleted the messagebase directory of all files
first. I created all needed directories and files. So here is a list of
how I started out before I invoked any of the mail tossers.
Delete all files from inbound directory
delete all files from messagebase directory
Delete all log files
delete all bad files and dupes
move *.pkt files to inbound
move empty messagebase files to messagebase directory
And here is the system I'm running on :
386/20 DX
DOS 5.0
640k cache with staged writes
4DOS 4.01
4 megs of ram with 1.5 ramdrive for swapping if needed
110 Meg ESDI 16ms access hard drive
This is my normal setup. But for this comparison, I disabled the
cache completely. The actual times are not important and should not
be taken as written in stone. What is important, is the ratio in the
different times. So if my setup shows something different, it will
depend on your setup. Using a cache should also speed up the tossing
time as well. It might be something like this :
mine 486/33 XT
Tosser A 20 secs 10 secs 40 secs
Tosser B 15 secs 7.5 secs 30 secs
As some commercials state, "Your mileage my vary".
I used 31 *.PKT files from 4 different networks. They were all
different sizes and contained 650 messages total. I'm not going to
put the logfile results in this article because some of them are
quite lengthy. It also would not serve much purpose. The times that I
have were created using the 4dos Timer command. I started the timer,
invoked the tosser, and then turned off the the timer. So the times
shown are actually a little longer than the actual times as reported
in the logfiles. I thought that if I did this for all of the tossers,
it would be a bit better for comparisons. Some logfiles showed start
and stop times but not how long they were active. And rather than
make a mistake in my calculations, I thought that I would let 4dos do
the arithmetic. I also show two results for the space usage. One is
space allocated and the other is space used by files.
Here is a list of the tossers that I tried and the versions :
Ezpoint 2.2 (Unique, points only)
Fastecho 1.20A (Hudson)
Fmail 0.92 (beta) (Hudson)
Freemail 1.00 (beta ?) (Hudson)
Gecho (beta) (Hudson)
Imail 1.21A (Hudson)
Ppoint 1.35 (Unique, points only)
Qecho 2.75 (Hudson
Qmail 1.30 (gamma) (Fido style)
Spoint 1.20 (Hudson)
Squish 1.01 (Squish)
Zztoss (beta) (Hudson)
And here is a table of the combined results :
Time Files Bytes Bytes Allocated
Ezpoint 2:58.13 28 607,520 632,832
Fastecho 0:28.29 9 928,664 937,984
Fmail 0:37.63 9 931,224 940,032
Freemail 2:13.79 9 991,896 1,001,472
Gecho 1:04.59 9 839,832 847,872
Imail 5:10.11 9 839,320 847,872
Ppoint 3:18.56 53 799,028 862,208
Qecho 2:24.45 9 812,696 821,248
Qmail 4:08.10 869 1,120,556 1,802,240
Spoint 1:01.30 9 928,664 937,984
Squish 2:54.50 81 930,492 1,005,568
Zztoss 1:30.03 7 840,042 845,824
Tossers from fastest to slowest :
1 Fastecho 0:28.29
2 Fmail 0:37.63
3 Spoint 1:01.30
4 Gecho 1:04.59
5 Zztoss 1:30.03
6 Freemail 2:13.79
7 Qecho 2:24.45
8 Squish 2:54.50
9 Ezpoint 2:58.13
10 Ppoint 3:18.56
11 Qmail 4:08.10
12 Imail 5:10.11
Disk Usage (File bytes) least to most :
1 Ezpoint 607,520
2 Ppoint 799,028
3 Qecho 812,696
4 Imail 839,320
5 Gecho 839,832
6 Zztoss 840,042
7.5 Fastecho 928,664
7.5 Spoint 928,664
9 Squish 930,492
10 Fmail 931,224
11 Freemail 991,896
12 Qmail 1,120,556
Disk Usage (Allocation bytes) least to most :
1 Ezpoint 632,832
2 Qecho 821,248
3 Zztoss 845,824
4.5 Imail 847,872
4.5 Gecho 847,872
6 Ppoint 862,208
7.5 Fastecho 937,984
7.5 Spoint 937,984
9 Fmail 940,032
10 Freemail 1,001,472
11 Squish 1,005,568
12 Qmail 1,802,240
Summary :
---------
This is not meant to be an endorsement of any tosser nor is it meant
to put down any tosser. It is not meant to be "The Definitive Test".
As I stated to begin with -- I set these tossers up according to how
I would use them. Everyone has different needs. Some of these tossers
will only work within their own enviornment, others may be used by lots
of other programs. Some of the tossers are only for point systems.
Others are Hudson style mail tossers and Squish has it's own format.
Qmail is a Fido style mail tosser. It is also up to the individual
person which format they prefer. Is speed is more important than
space used or the other way around ? What about cost ? Do you want
the messagebase in one file ? Or do you want the areas separated ?
Some of these tossers are in testing stages, some are fairly new,
others are well proven. And I found that some tossers are quite easy
to setup while others are a bit more complicated. Some have a setup
program while others require you to edit a configuration file.
I thought that I share some results with you about what I found
out when looking at these tossers. All of them work well and are
worth taking a look at.
Feel free to send comments to me :
Charles Buchanan 1:3812/10.6
===
Всяческих благ. Искренне Ваш, Vladimir Fyodorov, эсквайр.
... Похоже, я опять попал в оффтопик...
--- GoldED+/OSX 1.1.5-b20250401
* Origin: Esquire Station (2:6035/3.1)
