Where can I find help on the user interface for Reduce (Free CSL Version)? (Help F1 is not working in my installation.)
What operating system do you use? In my Debian 8.2 (and I guess also in
a recent Ubuntu) the key F1 works.
If you would fetch the source code of reduce you would find the manual
and more useful documents in the folder 'doc'.
In particular, are there ways to select a particular output, copy it, and paste it into some editor?
cut and paste from the menus is temporarily not working, at least for me
... in Linux the standard X select-and-paste works. If you wish to
select the output you should switch off the LaTeX output by
off fancy;
off nat;
then you have an output that can be used again as a reduce input. If you
wish a LaTeX source output, do
Thank you very much for your answer. My OS is Windows 7.
I did find a manual there and a manual online, but did not find answers there.
Using the switches does allow me now to select the output. However, after I select it and choose either Copy or Copy Text from the Edit menu, and then Control-V, nothing gets pasted, neither in the reduce window nor outside of it.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Thank you very much for your answer. My OS is Windows 7.
Then there is no X selection mechanism, as I told in my last email for
Linux.
I did find a manual there and a manual online, but did not find answers there.
At the moment the GUI is not documented, sorry.
Using the switches does allow me now to select the output. However, after I select it and choose either Copy or Copy Text from the Edit menu, and then Control-V, nothing gets pasted, neither in the reduce window nor outside of it.
There is a bug in cut-and-paste in the GUI, I think, maybe it will be
fixed soon. Other developers are more competent than me on this subject
and might discuss this matter. One thing that you might do is to run
reduce from a terminal (or console). Issue the command "redcsl -w" or
"redpsl" and you will have a text-mode reduce in a console with working
cut-and-paste from the Windows environment (ie, not dependent on the
reduce GUI).
Ages ago before the (GUI) CSL version of Reduce used display maths (ie
when it built up the display of formula using just one font and with
characters all the same size in a fixed-pitch grid) I had implemented the
ability to COPY from the output, and to PASTE from copied stuff back into
the input. When I had implemented that I made the COP operation extrat
material in HTML format in such a way that prompt strings were in a
different colour from previously typed user-input, and via that I was able
to arrange that if one selected then COPIED a multi-line segment of
previous input and PASTEd it back the prompt marks in what you had COPIED
did not interfere. When I implemented the "fancy" style display that put
thinsg where there was not any (simple) textual form available for
COPYing, and a thinking about a graphical form is messy and almost
certainly not what anybody wants. The result was that support for COPY and
PASTE went on the (long) "to do" list that I have with a range of real
uncertainties in my mind about what the best behaviour could be. For
instance if and when I re-introduce it I feel I ought to make it possible
for the user to select (and hence COPY) a fragment from within a bit of
displayed mathematics. That might sound sort of obvious but I do not think
that in detaiul it is. At a minimum consider a formula as simple as "a + b
* c" and consider whether the text string "a + b" should be selectable. It
is clear that "b * c" ought to be, but picking up just "a + b" runs
against the semantics. That issue of design has been part of what has
stalled me, as has the fact that evenn if I was certain I knew what to do
it would be a noticable amount of work (and probably re-work to adjust
interbals in messy ways, since tracking where bits of a displayed formula
appear on a screen is quite a lot of code). And then deciding what format
the user wants the COPIEd material to appear in and whether it is
re-inputable to REDUCE....
At present although returning to that is on my stack to look into again
at some stage I have a noticable number of other things I am trying to
work on - some of which I see as building foundations that could support a
new display engine that (ideally) would have editing designed in from the
start... So fixing this is something that I am not about to be able to do.
However I remind all thar Reduce is Open Source and so any volunteer who
was willing to get their hands dirty could have a go themselves. For
instance the relevant code will be in csl/fox/src/FXTerminal.cpp and
FXShowMath.cpp etc and while I might not (sanely) drop all my other
projects to divert to address that particular project myself it is
probably that if somebody else was making a plausible and serious job of
improving things there I would be able to advise and help them.
WRT the current practical reality I have a could of fall-back suggestions.
For arranging that you can review everything after your session I often
launch reduce with the option "-L logfile" so that a transcript of what I
have done is put in that file. If one HAS to launch Reduce by
double-clicking on an icon this may invove settting command-line parametsr
in a shortcut to it. In general I live sitting at a cygwin command line so
I just ensure that the relevant directoty is on my PATH then go
redcsl -L logfile.log
The other suggestion that supports copy and paste a bit is
redcsl -w [optionally -L logfile.log]
which runs Reduce in console mode, and then the standard
Cygwin/Xterm/Windows (or whatever console you are running in) copy & paste
facilities should be available. But that tends to mean that if you copy
several lines of previous input at once then the prompts are caught in
what you copy and pasting it back may be disrupted by that!
I suppose I should also note that at one stage some people got an
interface to TeXmacs going. I have not tried that for AGES and when I did
I only investigated for a short amount of time - and if TeXmacs has
changed since then unless somebody (who will not be me!) has had the
enthusiasm to review and (if necessary) update the fils in
csl/cslbase/texmacs-plugin I can not tell you how well that works today!
And of course if you use that or any other "front end" then just what
happens is up to that front end...
You mention that Mathematica has all sorts of user-interface capabilities.
I accept that. I believe they will have had a substantial team working on
that side of things - and whether that or the algebra engine has consumed
the most work is not quite obvious to me. On that "you pays your money and
makes your choice". The current deal where you get a copy of Mathematice
if you run a Raspberry Pi is an AMAZINGLY good deal (subject to the
capacity and speed of the Raspberry pi and how elaborate the algebra that
interests you is). Otherwise I like the fact that REduce is cost free to
you. I like the fact that all the source is available to you (and
everybody else) so that in the limit if you want to check anything you can
pretty way all the way down). I like to hope that the responsiveness on
this list is good and can help people - and that many glitches get
addressed promptly (but copy and paste will not, at least by me), and I
believe that while there are some algebra areas where Mathematica wins
that there will be others where users have found Reduce serves them
better. This last can depend very much on the individual user and the
class of problems they are needing to solve...
With Reduce there can always me more to do, so gradually sucking in more
develpers is a good plan. Very often people have started that because they
have a personal project or need for something and after a while they
discover that the insides of Reduce is "justa pile of code" written over
the years by a large number of different individuals, and that they may be
able to contribute their fragments....
Hope this at least explains a bit where we are, gives some (slightly
clumsy) work-rounds and encourages you to think of Reduce as a project as
well as a product.
You also noted a "bad tags in GC" message. That looks like a glitch that
ought to be my job to investigate. Can you please send me a script that
re-creates it. That involves making certain I know exactly which version
of Reduce you were running. When I have a fix that will be checked back
into the subversion repository and you would then be able to get the
benefit my running "svn update" and re-building. Indeed one of the first
things I will be ccecking is whether the very most recent version still
exhibits the bad behaviour. I have been working on the low-level guts of
the system fairly intensively over Christmas and there will (I am certain)
have been times when there were broken revisions! So try sending (to just
me) exact details of the version of Reduce you are using plus the input
file you provide it with and a transcript of the log up to and including
the "bad tags" message (or a screenshot if that is the best you can manage
there) so I can see exactly how far things got. I am not certain if we
will then count as lucky if that lets me reproduce the crash or kbeing
lucky would be if it just worked for me!!!! But discovering which of those
apply is surely step 1.
Regards. Arthur
On Mon, 18 Jan 2016, Raffaele Vitolo wrote:
On 18/01/16 06:13, Iosif Pinelis wrote:
Thank you very much for your answer. My OS is Windows 7.
Then there is no X selection mechanism, as I told in my last email for
Linux.
I did find a manual there and a manual online, but did not find answers there.
At the moment the GUI is not documented, sorry.
Using the switches does allow me now to select the output. However, after I select it and choose either Copy or Copy Text from the Edit menu, and then Control-V, nothing gets pasted, neither in the reduce window nor outside of it.
There is a bug in cut-and-paste in the GUI, I think, maybe it will be
fixed soon. Other developers are more competent than me on this subject
and might discuss this matter. One thing that you might do is to run
reduce from a terminal (or console). Issue the command "redcsl -w" or
"redpsl" and you will have a text-mode reduce in a console with working
cut-and-paste from the Windows environment (ie, not dependent on the
reduce GUI).
Arthur, I am attaching a log file with "Bad tag bits in GC", as well as the corresponding screen shot of the Reduce session with the About Reduce information. Thank you very much again.
You mention that Mathematica has all sorts of user-interface
capabilities.
I accept that. I believe they will have had a substantial team working on
that side of things - and whether that or the algebra engine has consumed
the most work is not quite obvious to me. On that "you pays your money and
makes your choice". ... Otherwise I like the fact that REduce is cost free to
you. I like the fact that all the source is available to you (and
everybody else) so that in the limit if you want to check anything you can
pretty way all the way down). I like to hope that the responsiveness on
this list is good and can help people - and that many glitches get
addressed promptly (but copy and paste will not, at least by me), and I
believe that while there are some algebra areas where Mathematica wins
that there will be others where users have found Reduce serves them
better. This last can depend very much on the individual user and the
class of problems they are needing to solve...
Here I'd like to add that Reduce was (and still is to some extent) the
favourite tool of many researchers in mathematics, I know many of them.
As far as I'm concerned, I had the possibility to learn a lot from the
fact that the source code was open and the experienced developers were
always willing to help me.
On the side of the effectiveness of the algebra-engine my experience is
that in many basic operations with rational expressions (like sums or
products) Reduce can get an answer where Maple and/or Mathematica cannot.
On the other hand there are packages or more specific tasks where Maple
and/or Mathematica perform better. In my opinion this is mostly due to
the fact that some Reduce packages are no longer maintained by their
developer. Just as an example I had much better results in Maple with
the Jordan form of matrices with respect to the Reduce package normalform.
But in latest years there were constant (even if slow) improvements to a
lot of parts in Reduce, and that is a positive sign. If anybody needs
better performance from a specific package a post in the mailing list
can attract the attention of developers and eventually a fix or an
improvement.
Raf.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Very late comment: when you get it working the texmacs interface displays great! Unfortunately I am at Ubuntu mint 14.04 , and texmacs 1.07.18 ; because of a previous error (lack of correct reducerc?). If anybody wants me to try mint 16 and/or texmacs 1.99 let me know. Unfortunatelly, while texmacs exists and has a support network there seems to be confusion about it's future. In any case mint 16 does not have it on it's synaptic list.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
If you use the code for the texmacs plugin in generic/texmacs and follow
the instruction in the README file, there should be no problem with running
REDUCE within TeXmacs. You don't need a .reducerc for that.
Eberhard
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Thanks Eberhard! That works and I have updated my installation script. Two small details.
-- "redcsl --version " generates version information and quits. It would be nice for automated installation record generation if "redpsl --version" did the same.
-- While fresh svn downloads to a clean target directory seem to work consistantly and produce reproducable results; "svn update" seems to produce erratic file sizes in "1>,2>" file lengths for statements like "./configure --with-psl 1>./rec-lists/config-psl.rec 2>./rec-lists/config-psl.err
make clean
make 1>./rec-lists/make-psl.rec 2>./rec-lists/make-psl.err"
I haven't tracked the cause down but I think it's a lot better with the
"make clean"
inserted before making anything.
If you are interested I will try to track down the offending processes.
I make everything; is it necessary to do the base config --with-xxx, (make clean?), make ; before going into debianbuild and doing another make?
If you or anyone would like a copy of my script for installation and record keeping (to allow automated comparisons of various installs) let me know. It's not polished but is readable (and boring) in a linear fashon so items can be commented out or put in.
Ray
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I forgot to mention: the present working installation here is:
Ubuntu mint 16.04.
TeXmacs 1.99.5
reduce-algebra revision 3826
Both csl and psl produce good testall results with few (I guess expected) results; reported errors seem to be in things called "-shifts" and "-rounding".
I have all the records/files of results.
Ray
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hello,
Where can I find help on the user interface for Reduce (Free CSL Version)? (Help F1 is not working in my installation.)
In particular, are there ways to select a particular output, copy it, and paste it into some editor?
Thank you.
On 17/01/16 15:44, Iosif Pinelis wrote:
What operating system do you use? In my Debian 8.2 (and I guess also in
a recent Ubuntu) the key F1 works.
If you would fetch the source code of reduce you would find the manual
and more useful documents in the folder 'doc'.
cut and paste from the menus is temporarily not working, at least for me
... in Linux the standard X select-and-paste works. If you wish to
select the output you should switch off the LaTeX output by
off fancy;
off nat;
then you have an output that can be used again as a reduce input. If you
wish a LaTeX source output, do
load_package tri;
on tex;
Raf.
Thank you very much for your answer. My OS is Windows 7.
I did find a manual there and a manual online, but did not find answers there.
Using the switches does allow me now to select the output. However, after I select it and choose either Copy or Copy Text from the Edit menu, and then Control-V, nothing gets pasted, neither in the reduce window nor outside of it.
On 18/01/16 06:13, Iosif Pinelis wrote:
Then there is no X selection mechanism, as I told in my last email for
Linux.
At the moment the GUI is not documented, sorry.
There is a bug in cut-and-paste in the GUI, I think, maybe it will be
fixed soon. Other developers are more competent than me on this subject
and might discuss this matter. One thing that you might do is to run
reduce from a terminal (or console). Issue the command "redcsl -w" or
"redpsl" and you will have a text-mode reduce in a console with working
cut-and-paste from the Windows environment (ie, not dependent on the
reduce GUI).
Raf
Ages ago before the (GUI) CSL version of Reduce used display maths (ie
when it built up the display of formula using just one font and with
characters all the same size in a fixed-pitch grid) I had implemented the
ability to COPY from the output, and to PASTE from copied stuff back into
the input. When I had implemented that I made the COP operation extrat
material in HTML format in such a way that prompt strings were in a
different colour from previously typed user-input, and via that I was able
to arrange that if one selected then COPIED a multi-line segment of
previous input and PASTEd it back the prompt marks in what you had COPIED
did not interfere. When I implemented the "fancy" style display that put
thinsg where there was not any (simple) textual form available for
COPYing, and a thinking about a graphical form is messy and almost
certainly not what anybody wants. The result was that support for COPY and
PASTE went on the (long) "to do" list that I have with a range of real
uncertainties in my mind about what the best behaviour could be. For
instance if and when I re-introduce it I feel I ought to make it possible
for the user to select (and hence COPY) a fragment from within a bit of
displayed mathematics. That might sound sort of obvious but I do not think
that in detaiul it is. At a minimum consider a formula as simple as "a + b
* c" and consider whether the text string "a + b" should be selectable. It
is clear that "b * c" ought to be, but picking up just "a + b" runs
against the semantics. That issue of design has been part of what has
stalled me, as has the fact that evenn if I was certain I knew what to do
it would be a noticable amount of work (and probably re-work to adjust
interbals in messy ways, since tracking where bits of a displayed formula
appear on a screen is quite a lot of code). And then deciding what format
the user wants the COPIEd material to appear in and whether it is
re-inputable to REDUCE....
At present although returning to that is on my stack to look into again
at some stage I have a noticable number of other things I am trying to
work on - some of which I see as building foundations that could support a
new display engine that (ideally) would have editing designed in from the
start... So fixing this is something that I am not about to be able to do.
However I remind all thar Reduce is Open Source and so any volunteer who
was willing to get their hands dirty could have a go themselves. For
instance the relevant code will be in csl/fox/src/FXTerminal.cpp and
FXShowMath.cpp etc and while I might not (sanely) drop all my other
projects to divert to address that particular project myself it is
probably that if somebody else was making a plausible and serious job of
improving things there I would be able to advise and help them.
WRT the current practical reality I have a could of fall-back suggestions.
For arranging that you can review everything after your session I often
launch reduce with the option "-L logfile" so that a transcript of what I
have done is put in that file. If one HAS to launch Reduce by
double-clicking on an icon this may invove settting command-line parametsr
in a shortcut to it. In general I live sitting at a cygwin command line so
I just ensure that the relevant directoty is on my PATH then go
redcsl -L logfile.log
The other suggestion that supports copy and paste a bit is
redcsl -w [optionally -L logfile.log]
which runs Reduce in console mode, and then the standard
Cygwin/Xterm/Windows (or whatever console you are running in) copy & paste
facilities should be available. But that tends to mean that if you copy
several lines of previous input at once then the prompts are caught in
what you copy and pasting it back may be disrupted by that!
I suppose I should also note that at one stage some people got an
interface to TeXmacs going. I have not tried that for AGES and when I did
I only investigated for a short amount of time - and if TeXmacs has
changed since then unless somebody (who will not be me!) has had the
enthusiasm to review and (if necessary) update the fils in
csl/cslbase/texmacs-plugin I can not tell you how well that works today!
And of course if you use that or any other "front end" then just what
happens is up to that front end...
You mention that Mathematica has all sorts of user-interface capabilities.
I accept that. I believe they will have had a substantial team working on
that side of things - and whether that or the algebra engine has consumed
the most work is not quite obvious to me. On that "you pays your money and
makes your choice". The current deal where you get a copy of Mathematice
if you run a Raspberry Pi is an AMAZINGLY good deal (subject to the
capacity and speed of the Raspberry pi and how elaborate the algebra that
interests you is). Otherwise I like the fact that REduce is cost free to
you. I like the fact that all the source is available to you (and
everybody else) so that in the limit if you want to check anything you can
pretty way all the way down). I like to hope that the responsiveness on
this list is good and can help people - and that many glitches get
addressed promptly (but copy and paste will not, at least by me), and I
believe that while there are some algebra areas where Mathematica wins
that there will be others where users have found Reduce serves them
better. This last can depend very much on the individual user and the
class of problems they are needing to solve...
With Reduce there can always me more to do, so gradually sucking in more
develpers is a good plan. Very often people have started that because they
have a personal project or need for something and after a while they
discover that the insides of Reduce is "justa pile of code" written over
the years by a large number of different individuals, and that they may be
able to contribute their fragments....
Hope this at least explains a bit where we are, gives some (slightly
clumsy) work-rounds and encourages you to think of Reduce as a project as
well as a product.
You also noted a "bad tags in GC" message. That looks like a glitch that
ought to be my job to investigate. Can you please send me a script that
re-creates it. That involves making certain I know exactly which version
of Reduce you were running. When I have a fix that will be checked back
into the subversion repository and you would then be able to get the
benefit my running "svn update" and re-building. Indeed one of the first
things I will be ccecking is whether the very most recent version still
exhibits the bad behaviour. I have been working on the low-level guts of
the system fairly intensively over Christmas and there will (I am certain)
have been times when there were broken revisions! So try sending (to just
me) exact details of the version of Reduce you are using plus the input
file you provide it with and a transcript of the log up to and including
the "bad tags" message (or a screenshot if that is the best you can manage
there) so I can see exactly how far things got. I am not certain if we
will then count as lucky if that lets me reproduce the crash or kbeing
lucky would be if it just worked for me!!!! But discovering which of those
apply is surely step 1.
Regards. Arthur
On Mon, 18 Jan 2016, Raffaele Vitolo wrote:
Arthur, I am attaching a log file with "Bad tag bits in GC", as well as the corresponding screen shot of the Reduce session with the About Reduce information. Thank you very much again.
On 18/01/16 10:24, Arthur Norman wrote:
Here I'd like to add that Reduce was (and still is to some extent) the
favourite tool of many researchers in mathematics, I know many of them.
As far as I'm concerned, I had the possibility to learn a lot from the
fact that the source code was open and the experienced developers were
always willing to help me.
On the side of the effectiveness of the algebra-engine my experience is
that in many basic operations with rational expressions (like sums or
products) Reduce can get an answer where Maple and/or Mathematica cannot.
On the other hand there are packages or more specific tasks where Maple
and/or Mathematica perform better. In my opinion this is mostly due to
the fact that some Reduce packages are no longer maintained by their
developer. Just as an example I had much better results in Maple with
the Jordan form of matrices with respect to the Reduce package normalform.
But in latest years there were constant (even if slow) improvements to a
lot of parts in Reduce, and that is a positive sign. If anybody needs
better performance from a specific package a post in the mailing list
can attract the attention of developers and eventually a fix or an
improvement.
Raf.
Raffaele and Arthur, thank you very much for your answers.
Very late comment: when you get it working the texmacs interface displays great! Unfortunately I am at Ubuntu mint 14.04 , and texmacs 1.07.18 ; because of a previous error (lack of correct reducerc?). If anybody wants me to try mint 16 and/or texmacs 1.99 let me know. Unfortunatelly, while texmacs exists and has a support network there seems to be confusion about it's future. In any case mint 16 does not have it on it's synaptic list.
If you use the code for the texmacs plugin in generic/texmacs and follow
the instruction in the README file, there should be no problem with running
REDUCE within TeXmacs. You don't need a .reducerc for that.
Eberhard
Thanks Eberhard! That works and I have updated my installation script. Two small details.
-- "redcsl --version " generates version information and quits. It would be nice for automated installation record generation if "redpsl --version" did the same.
-- While fresh svn downloads to a clean target directory seem to work consistantly and produce reproducable results; "svn update" seems to produce erratic file sizes in "1>,2>" file lengths for statements like "./configure --with-psl 1>./rec-lists/config-psl.rec 2>./rec-lists/config-psl.err
make clean
make 1>./rec-lists/make-psl.rec 2>./rec-lists/make-psl.err"
I haven't tracked the cause down but I think it's a lot better with the
"make clean"
inserted before making anything.
If you are interested I will try to track down the offending processes.
I make everything; is it necessary to do the base config --with-xxx, (make clean?), make ; before going into debianbuild and doing another make?
If you or anyone would like a copy of my script for installation and record keeping (to allow automated comparisons of various installs) let me know. It's not polished but is readable (and boring) in a linear fashon so items can be commented out or put in.
Ray
I forgot to mention: the present working installation here is:
Ubuntu mint 16.04.
TeXmacs 1.99.5
reduce-algebra revision 3826
Both csl and psl produce good testall results with few (I guess expected) results; reported errors seem to be in things called "-shifts" and "-rounding".
I have all the records/files of results.
Ray