How does it compare to FireFox?
(Systems supported? Security? Ease of
use? Configurability?
I'm not interested in doing any detailed evaluations of browsers. Do
you get involved with the forums for K-M, Firefox and Moz? I've simply
long been interested in K-M as a light and fast alternative browser.
Since this new version is based on Moz 1.8a2 it's said to be the most
secure.
I suspect that it loads faster than Firefox. Mozilla has a nice
feature whereby you can leave a portion of it in memory for (much)
faster loading. Firefox and K-M don't have that feature, and it seems
K-M wouldn't need it
It loads very quickly on my PC.
Downloading? (Can it continue an interrupted
download where it left off? Does it download a file completely?[1]))
I have no problems at all with interrupted downloads with my DSL
service. I just did a quick comparison of downloading Moz 1.7.2 with
both Moz and K-M, and in both cases the file length is 12,043,984
bytes.
That's B7C6D0 in hexadecimal which indicates that the partial buffer of
data fetched at the end of the file is correctly written to disk.
Good. I have had problems with both IE's download routine and FireFox's
download manager. When an attempt is made to continue an interrupted
download later, you are asked if you want to overwrite the
previously-partially-downloaded file or rename the file you are currently
asking it to download (even though it's the same file and just more
complete). GetBot had an option to continue saving where you left off
but its way of implementing appeared to me to be to start downloading at
the beginning of the file all over again and just discard everything until
it reached the spot you left off previously and then start writing to
disk. I now use wget when I can't use lynx on my ISP's system.[1] And
the file truncation is also a problem. The file-truncation problem came
to light when a user here reported having problems with files downloaded
through his CCN (Chebucto Community Net) account. They were smaller than
files downloaded with a different machine and failed to install properly,
citing CRC errors. I now suspect that the two machines were running
different software to perform the downloads.
My interest is partly due to my being a volunteer Bob[2] at my
(non-profit, volunteer-run) ISP. I might end up having to support it
now that my ISP recommends that people avoid using Internet Explorer
and Outlook and Outlook Express as much as possible due to security
concerns. (And it's partially on-topic here as problems caused by
file-truncation could easily be falsely blamed on viruses.)
[1] Tests with downloading one file resulted in consistantly getting
three different file sizes depending on the application or browser
used for the download (repeated more than once for each browser or
application except IE which I only dared to try once):
1941504 bytes: -- Downloaded with Firefox 0.8
-- Downloaded with Firefox 0.9.3
-- Downloaded with Internet Explorer
-- Downloaded with PC-LYNX
1942590 bytes: -- Fetched with lynx on my ISP's Unix server and
downloaded with ZModem
-- Downloaded with wget
-- Downloaded with GetBot
-- Downloaded with links [sic] 0.98
1949953 bytes: -- Downloaded with links [sic] 0.83
Another interesting thing comes to light when two of the file sizes
are converted to hexadecimal:
Decimal Hexadecimal
1,942,590 1DA43E
1,941,504 1DA000
Hmm. Did you examine the downloads using a hex editor to try to find
what the differences are? Did the downloaded files all work in spite
of the differences?
Well, I didn't try running the suspected bad files. (It's the install
routine for a program and I don't want to install multiple copies.)
Truncated versions of a larger version of that file (also with "000"
as the last three digits of the hexadecimal download size) have been
reported by the user mentioned above as being corrupted when he tried to
run it but downloads through other channels were larger and ran OK. (I
suspect that the software distributor changed builds between the time the
user had problems and the time I downloaded the files to test them and the
new build was smaller.) He reported similar problems with a different
program file but I haven't found the download address for that one yet.[0]
After some hex-viewing, it appears as though the 1941504-byte files are
truncated by not having the last partial buffer of data being written to
disk. And the 1949953-byte files downloaded with links version 0.83 are
corrupted copies because the downloads were incorrectly done in ASCII mode
instead of binary and all of the 0Ah bytes in the files were falsely
assumed to be UNIX newlines and were changed to the two-byte sequence 0Dh,
0Ah (ie. line-feeds in the file were changed to carriage-return/line-feed
pairs even when they were not used as line feeds).
[0] The file I *did* find the download for was at:
http://download.jgsoft.com/editpad/SetupEditPadProDemo.exe
[1] Such as when the server at the other end responds so sluggishly
that the download time for lynx to fetch the file exceeds the ten
minutes and two-minute grace period for an idle session. Any keypress
that will satisfy the idle detector and stop me from being logged off
is also one that will tell lynx to cancel the download. Once the file
has been fetched by lynx, I can select a ZModem download and the
idle-detector will be disabled during the local download.
[2]
http://www.chebucto.ns.ca/~af380/atsr-faq.html#x1224x