Some comments regarding the translation of resources and www
Thanos Kakarountas
I have some comments regarding the new version based on the english
translation and some designs made with the new version.
Comments
-----------------
The multiplier does not operate this way. Carry in and Carry out are
always one bit, and that's what is expected since a multiplication of 2
inputs (let's assume non equal widths, e.g. m x n bits) will result a
product of (m+n) bits wide plus a carry out bit resulted by the final
addition stage (sum of products))
What is the tunnel component? I haven't seen nor used something like
this before.
Some comments on the properties strings are following:
For resources translation
std.properties
<segment 0265> Clear: 1 pins contents to 0 asynchronously
<segment 0265> Clear: when 1, resets contents to 0 asynchronously
<segment 0301> Enable: when not 0, output is selected input
<segment 0301> Enable: when not 0, output is the selected input
<segment 0334> Borrow In: if 1, 1 is deducted from the output
<segment 0334> Borrow In: if 1, output is decreased by 1
<segment 0338> Output: the products of the inputs, plus the carry in
<segment 0338> Output: the product of the inputs, plus the carry in
<segment 0340> Carry Out: the upper bits of the product
<segment 0340> Carry Out: the upper bit (MSB) of the product
<segment 0366> Roll Left
<segment 0366> Rotate Left
<segment 0367> Roll Right
<segment 0367> Rotate Right
<segment 0373> Output: the number of 1 bits in the inputs
<segment 0373> Output: the amount of 1's in the input
<segment 0373> Output: the number of 1 bits in the inputs
isn't this the same as the above?
for www translation
download.html
<segment 0020> You will three choices of which release to download.
<segment 0020> You have three choices of which release to download.
<segment 0039> From the dropdown menu labeled Language, select Russian
language.
shouldn't that be: <segment 0039> From the dropdown menu labeled
Language, select English language.
----------------
End of Comments
I like very much the new version. Congratulations
C Burch
months, but it's good to have you back! I'm happy to hear that you're
about finished with the translation. As Ilia said, though, you haven't
yet been successful with uploading it to the Subversion repository.
You can browse to
http://circuit.svn.sourceforge.net/viewvc/circuit/trunk/src/resources/logisim/el/
to check on what has made it to the repository.
Once you successfully upload the version, I'm hoping to build a
fragile release so that you can see how the translation will appear to
end users.
I appreciate the suggestions. I agree that most of your rewordings are
definite improvements, and I'll incorporate most of them into the
upcoming version. You mention that there seem to be two strings
reading "Output: the number of 1 bits in the input"; the difference is
that the other one changes the final word to its plural, "...the
inputs".
On the multiplier, I think you are suggesting that the component
itself is misdesigned. The idea of the current design is that if, say,
you wanted to multiply a 32-bit value X by a 4-bit value Y, you might
do this using eight 4-bit multipliers cascaded together: Each of the
eight multipliers would get the same 4-bit Y value and a different
group of 4 bits from X, with the carry-out from one multiplier being
fed into the carry-in of the next one. The carry from one multiplier
to another could be as long as 4 bits.
I'm not sure how you are suggesting a multiplier *ought* to work - if
you get two four-bit inputs, plus a 1-bit carry-in input, it sounds
like you'd want an 8-bit output. And that's fine, but there wouldn't
be any possibility of a carry-out: The maximum multiplication result
would be 15 * 15 + 1, which fits into 8 bits. In any case, Logisim's
multiplier has worked this way since 2005, so I can't simply
completely discard the existing design - though it could be possible
for me to add in an option for the multiplier to work with 1-bit
carry-out - though, like I say, I don't yet see why somebody would
want that.
The "tunnel" is essentially a way to connect different points in a
circuit without having a wire between them. It's not strictly
necessary - you could simply draw wires between the two points - but
sometimes (I guess) it looks nicer not to have lines winding from one
corner of the layout to another. For my part, it feels very much like
a GOTO statement, so I don't use them. But some users have found it a
very important feature - and Georgia Tech was going so far as to fork
their own version of Logisim so they could have tunnels (among some
other features). Rather than see Logisim become forked, I integrated
the feature into the version I'm maintaining. I'd admit, though, that
"tunnel" may not be a mainstream word for the concept - that's
actually the word that Georgia Tech used in their version, and I just
kept it. From my point of view, it sort of makes sense, since it feels
sort of like there's a tunnel buried underneath the circuit's main
level, through which the wires travel.
Anyway, if you don't yet understand tunnels, I'd suggest that you
build the circuit diagrammed in the documentation
(http://ozark.hendrix.edu/~burch/logisim/docs/2.6.0/en/libs/base/tunnel.html)
and try different inputs.
-Carl
> --
> You received this message because you are subscribed to the Google Groups
> "Logisim" group.
> To post to this group, send email to log...@googlegroups.com.
> To unsubscribe from this group, send email to
> logisim+u...@googlegroups.com.
> For more options, visit this group at
> http://groups.google.com/group/logisim?hl=en.
>
>
Thanos Kakarountas
excuse me for not noticing your e-mail earlier. As you have probably
seen in an earlier e-mail, I have eventually uploaded the translation.
Thank you also for the tip how to check the repository. I had completely
forgot it!
I will start my comments/reply from the end of your e-mail.
'Tunnel' is a great idea! I haven't seen something like this before in a
commercial EDA tool and I couldn't think of something relevant. In fact
I had once created a script for a Mentor Graphics tool to do exactly the
same thing. It really helps designers not to miss the goal, abstracted
by the a of wires!
Regarding the multiplier, it would be nice to follow the typical
multiplier designs found in the technical literature (like Braun, or
Wallace, or Booth). Of course, I do not mean to design the exact
structure (although such a library would be very useful to the
students), but at least to follow the same output(s). Indeed a carry-out
doesn't have any meaning if we already have the bits that we already
need (i.e. for a mxn multiplication, the product fits in a m+n bits
word). However, if you check the following links (randomly selected),
you will see that what I was referring to is to the carry-out of the
last addition stage that corresponds to the Most Significant Bit (MSB)
of the product. Excuse me if I wasn't analytical in the earlier e-mail.
links:
www.wlu.ca/science/physcomp/.../pc319/.../Combinational%20Multiplier.pdf
www.writphotec.com/mano4/.../Mulitpliers_Dividers_supp4.pdf
There is no need to design from scratch or replace the already existing
multiplier component. Your suggestion (for configuring the carry-out) is
better. Of course the whole thing is minor, concerning the development
of Logisim. But it had to ask, since I haven't seen something like this
before, neither in the textbooks I teach to my courses, nor used it any
of my commercial designs.
Yes you are right about the two strings. Didn't notice it!
Thanos