New Version of SDEval and ACA'2015

[Albert Heinle]

Dear Symbolic-Data team,

I herewith announce that we, the SDEval team, have added several new
features in our SDEval toolbox. Please find details in the pull
request sent on GitHub yesterday
(https://github.com/symbolicdata/symbolicdata/pull/4). Furthermore, I
have recently presented SDEval in the context of a lab-meeting, and
recorded the presentation. I intended this as a tutorial for
interested users, and it can be viewed at
https://www.youtube.com/watch?v=CctmrfisZso.
We should also include that link in the SD Wiki.

It is of importance that this pull request is dealt with as soon as
possible. The reason is that, since the last update of SDEval on the
official Symbolic-Data repository, there apparently
have been changes made in the folder structure of the Symbolic-Data
project, which makes the currently included SDEval scripts ask the user
constantly for the respective locations.

Another point: I recently got an invitation to the "Applications of
Computer Algebra 2015" conference (ACA'2015), to present SDEval. It is
this year in Kalamata, Greece, and is promising to become a great
conference. Unfortunately, I will not be able to make it. Right now, I
am searching for a replacement for me, and decided to ask in this
mailing-list. Would there be anybody here willing to go there and give
this presentation? We can even shift the focus a little bit and talk
more about Symbolic-Data than SDEval, if it makes certain speakers
more comfortable (In fact, I think that also talking about the
computer algebra social network as Hans-Gert presented December 2012 would be
interesting.). If you are interested, let me know
by writing me a private message and we can arrange details. Papers on
presentations can later be published in the "Springer Proceedings in
Mathematics and Statistics". Find more information on
http://www.singacom.uva.es/ACA2015.

This is all from me right now.

All the best and have a great week.

Albert

[hgg]

Am 11.02.2015 um 19:07 schrieb Albert Heinle:
> It is of importance that this pull request is dealt with as soon as
> possible. The reason is that, since the last update of SDEval on the
> official Symbolic-Data repository, there apparently
> have been changes made in the folder structure of the Symbolic-Data
> project, which makes the currently included SDEval scripts ask the user
> constantly for the respective locations.

Hi Albert!

Nice to hear about such unexpected progress of the SD project. I have
some trouble to understand how things are related.

As explained at http://wiki.symbolicdata.org/Using.Git we run (for a
long time only) two branches 'master' and 'develop' at the SD master
repo https://github.com/symbolicdata/symbolicdata. To add a
contribution we advice there the following

* clone the develop branch of the public master repo to a private
account (best, if also at github),
* start a feature branch from develop on your blessed repo
* do valuable work
* issue a pull request to the Integration Manager Ralf Hemmecke ...

So it would help if you rebase your contribution first onto
symbolicdata/symbolicdata/develop. As far as I see you host also the
obsolete for a long time branch old-master as ioah86/old-master and a
very early stage of my hg-graebe/symbolicdata/workbench as
ioah86/develop. May be its time to get your repo in better sync with the
official repo structure first?

@Ralf: Please have a look at the develop branch of my
https://github.com/hg-graebe/symbolicdata repo. The last sync of
hg-graebe/symbolicdata/develop with symbolicdata/symbolicdata/develop
was half a year ago. Meanwhile I added the paper sources and some more
Draft work on RDF data to develop. I think that it is time to merge that
to symbolicdata/symbolicdata/develop, too.

Our policy was to have a LST version on the master branch. The last
release is from Sep 2013, so it may be a good idea to use the
opportunity offered by Albert to start to compile a new stable version?
From my point of view only the content of the data/Drafts subdir should
not yet moved to the master branch.

Best regards,
Hans-Gert

--

apl. Prof. Dr. Hans-Gert Graebe, Inst. Informatik, Univ. Leipzig
postal address: Postfach 100920, D-04009 Leipzig
Hausanschrift: Augustusplatz 10, 04109 Leipzig, Raum P-633
tel. : +49-341-97-32248
email: gra...@informatik.uni-leipzig.de
Home Page: http://www.informatik.uni-leipzig.de/~graebe

[Albert Heinle]

Hi Hans-Gert,

Ah, I knew I was overseeing something. Right, the development branch. Okay, I will get everything in sync and renew the pull request once it is done (should not take more than two weeks). I will for now withdraw the one that I just posted.

All the best,

Albert Heinle (albert...@googlemail.com)
47-20 Paulander Drive,
N2M-5L3, Kitchener, ON
Canada

> --
> You received this message because you are subscribed to the Google Groups "symbolicdata" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to symbolicdata...@googlegroups.com.
> To post to this group, send email to symbol...@googlegroups.com.
> Visit this group at http://groups.google.com/group/symbolicdata.
> For more options, visit https://groups.google.com/d/optout.

[hgg]

Am 12.02.2015 um 15:24 schrieb Albert Heinle:
> Hi Hans-Gert,
>
> Ah, I knew I was overseeing something. Right, the development branch. Okay, I will get everything in sync and renew the pull request once it is done (should not take more than two weeks). I will for now withdraw the one that I just posted.
>
> All the best,
>
> Albert Heinle

Hi Albert!

I merged your branch with my own develop (ahead of symbolicdata/develop
by 18 commits) and observed only minor problems:

> graebe@graebelaptop:~/git/SD/symbolicdata>git merge ioah86/master
> automatischer Merge von src/sdeval/ctc.py
> KONFLIKT (Inhalt): Merge-Konflikt in src/sdeval/ctc.py
> automatischer Merge von src/sdeval/create_tasks_gui.py
> KONFLIKT (Inhalt): Merge-Konflikt in src/sdeval/create_tasks_gui.py
> KONFLIKT (ändern/löschen): src/sdeval/create_tasks.py gelöscht in ioah86/master und geändert in HEAD. Stand HEAD von src/sdeval/create_tasks.py wurde im Arbeitsbereich gelassen.
> automatischer Merge von src/sdeval/classes/templates/comp/FA_Q_dp/Singular/template.py
> KONFLIKT (Inhalt): Merge-Konflikt in src/sdeval/classes/templates/comp/FA_Q_dp/Singular/template.py
> automatischer Merge von src/sdeval/classes/probleminstances/FreeAlgebrasFromXMLBuilder.py
> KONFLIKT (Inhalt): Merge-Konflikt in src/sdeval/classes/probleminstances/FreeAlgebrasFromXMLBuilder.py
> automatischer Merge von src/sdeval/classes/probleminstances/FreeAlgebras.py
> automatischer Merge von src/sdeval/classes/exceptions/XMLResourcesInvalidPath.py
> automatischer Merge von src/sdeval/classes/computationproblems/FA_Q_dp.py
> automatischer Merge von src/sdeval/classes/XMLResources.py
> KONFLIKT (Inhalt): Merge-Konflikt in src/sdeval/classes/XMLResources.py
> Automatischer Merge fehlgeschlagen; beheben Sie die Konflikte und committen Sie dann das Ergebnis.

So the merge task looks not very complicated.

[Ralf Hemmecke]

> @Ralf: Please have a look at the develop branch of my
> https://github.com/hg-graebe/symbolicdata repo. The last sync of
> hg-graebe/symbolicdata/develop with symbolicdata/symbolicdata/develop
> was half a year ago.

I've merged your develop branch and pushed to the official repo.

> Our policy was to have a LST version on the master branch. The last
> release is from Sep 2013, so it may be a good idea to use the
> opportunity offered by Albert to start to compile a new stable version?

OK.

> From my point of view only the content of the data/Drafts subdir should
> not yet moved to the master branch.

Uhm, what does that mean?

Your develop branch contains commits that include files in data/Drafts
with files that are not there. In particular data/RDFData/People.ttl.

https://github.com/symbolicdata/symbolicdata/commit/2939199f8ca3ce81a284cc91c55802059a09ace4

Of course, I can run "git filter-branch" and change (split) those
commits, but that will probably also have influence on the develop branch.

Ralf

[hgg]

Am 12.02.2015 um 22:40 schrieb Ralf Hemmecke:
>> > From my point of view only the content of the data/Drafts subdir should
>> >not yet moved to the master branch.
> Uhm, what does that mean?
>
> Your develop branch contains commits that include files in data/Drafts
> with files that are not there. In particular data/RDFData/People.ttl.
>
> https://github.com/symbolicdata/symbolicdata/commit/2939199f8ca3ce81a284cc91c55802059a09ace4
>
> Of course, I can run "git filter-branch" and change (split) those
> commits, but that will probably also have influence on the develop branch.

Okay, probably a bad idea. I just aligned the develop branch with my
workbench branch (concerning mainly a current state of the publication
section) - so please pull that also to the upstream develop branch.

My philosophy so far was to store regularly dumps of the RDF data on my
develop branch, hence new RDF data drafts should also reside there. This
is part of the current Backup concept as explained at
http://wiki.symbolicdata.org/Backups - weekly dumps of the RDF Data are
extracted directly from the Virtuoso store, monthly (archived) dumps of
both the RDF and XML Data are collected from a copy of the develop
branch of my SD repo fork at https://github.com/hg-graebe/symbolicdata.

So let's merge also the Drafts dir to the master branch, at least for now.

[Ralf Hemmecke]

Hi Albert,

On 02/11/2015 07:07 PM, Albert Heinle wrote:
> I intended this as a tutorial for
> interested users, and it can be viewed at
> https://www.youtube.com/watch?v=CctmrfisZso.
> We should also include that link in the SD Wiki.

Yes, I'm definitly for such a link.
Congratulations, Albert for this video.
I haven't yet deeply looked through your code, but from what I saw in
the video, it looks like quite a flexible an extensible design. It
somehow made me want to add FriCAS as a CAS to your options.

> It is of importance that this pull request is dealt with as soon as
> possible. The reason is that, since the last update of SDEval on the
> official Symbolic-Data repository, there apparently
> have been changes made in the folder structure of the Symbolic-Data
> project, which makes the currently included SDEval scripts ask the user
> constantly for the respective locations.

Well, that's somehow bad. But the problem here is probably that there is
no "make check" that tells me whether all the things (in a SD release)
are consistent. Would you be able to provide a few consistency checks
for SDEval? Bothering users with unnecessary questions is not a good way
to increase the wide-spead use of SD.

I'm waiting for your new pull request. I think, after I've merged it to
the official "develop", we can check and should probably advance master
to agree with develop (thus creating a new release).

Ralf

[hgg]

Am 16.02.2015 um 21:58 schrieb Ralf Hemmecke:
> I intended this as a tutorial for
>interested users, and it can be viewed at
>https://www.youtube.com/watch?v=CctmrfisZso.
>We should also include that link in the SD Wiki.

I just put the link on the page http://wiki.symbolicdata.org/SDEval.

Would it be a good idea to refactor that page, too?

hgg

[Albert Heinle]

Hello everyone

> On Feb 16, 2015, at 3:58 PM, Ralf Hemmecke <ra...@hemmecke.org> wrote:
>
> Hi Albert,
>
> On 02/11/2015 07:07 PM, Albert Heinle wrote:
>> I intended this as a tutorial for
>> interested users, and it can be viewed at
>> https://www.youtube.com/watch?v=CctmrfisZso.
>> We should also include that link in the SD Wiki.
>
> Yes, I'm definitly for such a link.
> Congratulations, Albert for this video.
> I haven't yet deeply looked through your code, but from what I saw in
> the video, it looks like quite a flexible an extensible design. It
> somehow made me want to add FriCAS as a CAS to your options.

Sure, that is possible. You can either try to do it yourself, given the instructions in the video and/or the documentation, or you can send me a certain code snippet and I enter it quickly.

>
>> It is of importance that this pull request is dealt with as soon as
>> possible. The reason is that, since the last update of SDEval on the
>> official Symbolic-Data repository, there apparently
>> have been changes made in the folder structure of the Symbolic-Data
>> project, which makes the currently included SDEval scripts ask the user
>> constantly for the respective locations.
>
> Well, that's somehow bad. But the problem here is probably that there is
> no "make check" that tells me whether all the things (in a SD release)
> are consistent. Would you be able to provide a few consistency checks
> for SDEval? Bothering users with unnecessary questions is not a good way
> to increase the wide-spead use of SD.

Yup, with the newest version, I have put A LOT OF TESTS in it. I concentrated mainly on patching everything with tests so I can safely extend the toolbox with features without destroying anything. For future consistency checks, you only have to run
$> nosetests -v
inside the sdeval folder, and if it says OK in the end, all the tests have passed and consistency is there. At the present moment, this is the case on my machine. FYI, if not yet known: nose is a very handy testing tool for Python code. For more information visit https://nose.readthedocs.org/en/latest/

>
> I'm waiting for your new pull request. I think, after I've merged it to
> the official "develop", we can check and should probably advance master
> to agree with develop (thus creating a new release).

As you probably already saw, I have created a new pull request to be in sync with the develop branch. Have fun pulling it, and let me know if there are any problems.

Sorry for taking so long, but last week, I had my presentation of my Ph.D research proposal, and that consumed, together with marking duties, almost all my time.

All the best,

Albert

>
> Ralf

[Albert Heinle]

Yes, I would agree that we should refactor that page. Give me a bit, and I will provide you with some text and some graphics you can embed into the wiki page.

All the best,

Albert

[Ralf Hemmecke]

Hi Albert,

as usual, first there will always be criticism. ;)

I'm a bit disappointed by your style of commit messages.
Have you tried "gitk" on your develop branch?

In th Graph, I see a lot of bullets that are just labelled by [UPDATE].
Although, in your full commit message you are a bit more verbose, your
format doesn't follow the suggestions by Junio C Hamano (main developer
of git).

http://doc.norang.ca/git-howto.html#sec-2

I don't expect that you are going to change that now, but please keep
that commit message convention in mind for future commits.

Ralf

[Ralf Hemmecke]

Hi Albert,

More comments to SDEval...

1)
I had to issue

apt-get install python-tk

before I could run "python create_tasks_gui.py". You should state in the
documentation what the prerequisites of sdeval are.

2)
The colors on the html page (proceedings.html), in particular green and
red are too heavy. I can hardly read the actual text. If the text color
is black, it would probably help if green and red were a bit lighter.

3)
If I press Ctrl-C to abort "python runTasks.py", the proceedings.html is
still telling me "RUNNING" for one task and "WAITING" for some others.
Would you be able to catch Ctrl-C and update to "ABORTED"?

4)
When I ran "python create_tasks_gui.py" for the second time, I again had
to enter the paths for my computer algebra systems and the time command.
I would have expected that this gets stored somewhere and offered (i.e.
pre-filled fields) when I see the form so that I would only have to
select the CAS and type in the new folder name and press ENTER.

5)
I don't see an order on the examples. Maybe at least an option (sorting
by example name) would be good. Perhaps also sorted by number of
variables or (expected) dimension of the problem.

6)
For some strange reason, I don't see anthing in the preview pane when I
select the (Gröbner) examples.

Ralf

[Albert Heinle]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi Ralf,

thank you for the valuable comments.

ad 1) That is interesting. Python-TK was always installed on all
machines I had (Ubuntu's and Mac's). What OS are you using? I agree,
we should state this in the documentation, if it is not standard. Good
catch

ad 2)
This is something that I already thought about. I have tentatively
created a css-file that gets copied always. But there is not much
content yet, and the interaction with the HTML is not yet at a good
state. It is on my todo, to make it look better :-)

ad 3) Puh, that is interesting. I personally have not given a thought
yet about the aborting option. The general philosophy is that you let
all of them run and early terminate some of the cas's in between in
case they take too long. If you abort early, you should know how to
interpret your data. But you are right, aborting might be an
interesting option to think about. I will add it to the TODO-list.

ad 4) GOOD IDEA. Never thought of this! I was thinking to redesign
that step anyway, and that would be cool to have.

ad 5) Yes, I agree... some sorting would be good.

ad 6) Known to me; when I fix some stuff around the GUI, I will
address this.

Thank you for the input. These are some good points. I will address
them as soon as possible.

All the best,

Albert

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
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=EqKE
-----END PGP SIGNATURE-----

[Albert Heinle]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi Ralf,

On 17.02.2015 17:14, Ralf Hemmecke wrote:
> Hi Albert,
>
> as usual, first there will always be criticism. ;)
>

And this is good :-) Constructional criticism makes projects better :-)

> I'm a bit disappointed by your style of commit messages. Have you
> tried "gitk" on your develop branch?
>
> In th Graph, I see a lot of bullets that are just labelled by
> [UPDATE]. Although, in your full commit message you are a bit more
> verbose, your format doesn't follow the suggestions by Junio C
> Hamano (main developer of git).
>
> http://doc.norang.ca/git-howto.html#sec-2

Didn't know that one. I will look into that. I personally do it "my
way" (which is actually learned during a project I participated in at
University), because then I can directly see debugs/updates on the
line. But okay, I am young, I can still learn ;-)

>
> I don't expect that you are going to change that now, but please
> keep that commit message convention in mind for future commits.

I will try.

All the best,

Albert

>
> Ralf
>

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
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=51it
-----END PGP SIGNATURE-----

[Ralf Hemmecke]

7)
SDEval's runTasks.py seems to only start the tasks in a serial order (no
parallel execution). That might be on purpose, but it would be good if
that is properly documented in the sdeval documentation. I don't see it
in sdeval-quick-manual.rst.

Ralf

[Ralf Hemmecke]

On 02/18/2015 02:34 AM, Albert Heinle wrote:
> ad 1) That is interesting. Python-TK was always installed on all
> machines I had (Ubuntu's and Mac's).

I use Ubuntu 14.10, but usually only install minimalistic.
Maybe my problem was also connected to starting python from a
virtualenv. Anyway apt-get install python-tk solved that problem.
Unfortunately, I couldn't find quickly, what to install via pip to get
the same functionality, so I ended up with apt-get.

Ralf

[Ralf Hemmecke]

> ad 3) Puh, that is interesting. I personally have not given a thought
> yet about the aborting option. The general philosophy is that you let
> all of them run and early terminate some of the cas's in between in
> case they take too long. If you abort early, you should know how to
> interpret your data. But you are right, aborting might be an
> interesting option to think about. I will add it to the TODO-list.

A change is not really necessary. I just thought, that it would be
better, if the status of the computation is properly documented. And
since you have this "proceedings.html" file...

If you decide to allow parallel benchmarks that such a status page is
probably difficult to implement. But since one usually expects to run
properly through all tasks without aborting any, I don't think you
should invest much time into this.

Aborting programs via the kill command is OK for the people who know how
to do this, but it's out of reach for people that don't know about
process numbers and such. BTW, does sdeval run on all operating systems?

Would it make sense to provide a command "abort_current_task" or
something like this?

Maybe, it's also worth thinking about adding the current running time to
the proceedings page (i.e. update it every minute) so that at least it's
approximately shown how long the current task is already running.

Ralf

[Albert Heinle]

Parallel execution is possible since the new version (one of the new features). Just add the flag -j and the number of tasks you want to run in parallel.

e.g.
$>python runTasks.py -j8

will run up to 8 CASs in parallel.

Albert Heinle (albert...@googlemail.com)
47-20 Paulander Drive,
N2M-5L3, Kitchener, ON
Canada

[Albert Heinle]

Oh, I see… Good to know. I will put this in the readme of my develop-branch today.

Albert Heinle (albert...@googlemail.com)
47-20 Paulander Drive,
N2M-5L3, Kitchener, ON
Canada

[Albert Heinle]

Ralf, you are speaking my language here ;-)

I fell I could do this abort-feature, and probably will at some point, but I will put it for now towards low-priority.

About the killing-business (this is the moment when NSA starts to listen ;-)): Yes, I totally agree. The manual killing is something that is a good solution, but for the non-expert-user, we might need something better in the future. I will think about a script to interact with runTasks.py in between.

About the current running-time on proceedings: Good idea. I was not thinking about that, as we have “top” to see things like that, and a maximum time the user can provide. I will keep that in mind for the future.

And the OS-question: As mentioned in the documentation, the create-tasks part should run on all possible OSs, the runTasks.py needs a Unix-like environment (Gnu/Linux, Mac OS). The reason is because I use “fork" in between to call the CAS’s and to be able to monitor them in parallel.
Our original design idea with Viktor and Benjamin was, that anybody can “create” tasks on their local machines as much as they want, but for the “running", people usually upload everything on a compute-server, which has usually a linux-based OS installed. If at some point a lot of Windows-users are coming to me and complaining that they can not use it on their Windows-servers, I might consider making a work-around; but for now, we are good I think.

Thanks again for the valuable feedback.

Albert Heinle (albert...@googlemail.com)
47-20 Paulander Drive,
N2M-5L3, Kitchener, ON
Canada

[Albert Heinle]

Hello Hans-Gert,

I have put some work into putting together a logo and some contents for that site. Please find it attached.

Let me know if there is anything you would like to change there.

All the best,

[Albert Heinle]

I have been informed, that .svg images (the format of the logo) are not necessarily uniformly displayed in all browsers.

In that case, I have converted the logo into the .png format. You can find it at https://cs.uwaterloo.ca/~aheinle/pictures/software/sdeval_logo.png

All the best,

Albert Heinle (albert...@googlemail.com)
47-20 Paulander Drive,
N2M-5L3, Kitchener, ON
Canada

> <sdeval_website.zip>

>
> Albert Heinle (albert...@googlemail.com)
> 47-20 Paulander Drive,
> N2M-5L3, Kitchener, ON
> Canada
>

>> On Feb 17, 2015, at 4:23 AM, hgg <h...@hg-graebe.de> wrote:
>>

[hgg]

Am 24.02.2015 um 16:50 schrieb Albert Heinle:
> I have put some work into putting together a logo and some contents for that site. Please find it attached.
> Let me know if there is anything you would like to change there.

I just put the description on the wiki page
http://wiki.symbolicdata.org/SDEval. The svg image worked fine. Please
have a look at the page. It would be helpful to have more Wiki editors.

For the ACA paper I had not yet time to think on it, but since it is in
the Session on Human-Computer Algebra Interaction, so I would probably
prefer to concentrate on CASN aspects.

Also I have no idea about the conference venue an SDEval poster may be
appropriate. How do you think about that?

Best regards, Hans-Gert

[Albert Heinle]

Hi Hans-Gert,

I just checked it: With Firefox on a Mac, the SVG-version cuts part of the “l” of SDEval out. Please use the PNG version I sent.

To the topic of Wiki-Editors: I could become one for the SDEval part, then you don’t need to make all the changes.

And ad ACA: Last year, I did not see any posters. I will ask Ilias if anything is different this year, and let you know. The CASN aspects are a good idea; I just wanted to offer some help in case you add SD to you talk.

All the best,

Albert Heinle (albert...@googlemail.com)
47-20 Paulander Drive,
N2M-5L3, Kitchener, ON
Canada

[Ralf Hemmecke]

On 03/09/2015 12:42 PM, Albert Heinle wrote:
> I just checked it: With Firefox on a Mac, the SVG-version cuts part
> of the “l” of SDEval out. Please use the PNG version I sent.

Albert, maybe it's an issue with the font. I've seen that in some of my
SVGs (generated with inkscape). After translating the characters to path
(see inkscape Path -> Object to Path (or Stroke-to-Path) there is no
relation to any font left so that no font-replacement will happen on
different machines.

Ralf

[Albert Heinle]

Oh, I did not have yet too much expertise with Inkskape (this logo was created after watching a bunch of video-tutorials ;-)). I will give it a try soon. In the meantime, the PNG is the image of choice.

Thanks,

Albert Heinle (albert...@googlemail.com)
47-20 Paulander Drive,
N2M-5L3, Kitchener, ON
Canada

[hgg]

Dear SD followers!

Am 08.03.2015 um 14:14 schrieb hgg:
> For the ACA paper I had not yet time to think on it, but since it is in
> the Session on Human-Computer Algebra Interaction, so I would probably
> prefer to concentrate on CASN aspects.

Albert proposed to help with some stuff on SDEval, but I think this is
inappropriate for that special session http://minimair.org/hcai2015/. I
prepared an abstract and submitted it as a first version. On 2015-03-30
Manfred Minimair as session chair asked for a submission until
2015-04-10, although the official deadline is 2015-05-15, hence there
was no possibility to discuss the submission in detail. But I think we
can submit a second version of the abstract if necessary.

Please have a look at http://wiki.symbolicdata.org/Publications, in
particular at http://symbolicdata.org/Papers/aca15-abstract.pdf, and
don't hesitate to comment all that (including my bad english).

Best regards,
hgg

[Albert Heinle]

Hello Hans-Gert,

I read through the abstract, and it was a good piece of work. I just have one question: The graph in the end, i.e. figure 1: Did you make it, or is it from one of your cited papers? You should add this as caption. Please let me know, because I would like to use it for an upcoming presentation, as it is quite a nice graph.

All the best,

Albert Heinle (albert...@googlemail.com)
47-20 Paulander Drive,
N2M-5L3, Kitchener, ON
Canada

[hgg]

Am 17.04.2015 um 14:35 schrieb Albert Heinle:
> I read through the abstract, and it was a good piece of work. I just have one question: The graph in the end, i.e. figure 1: Did you make it, or is it from one of your cited papers? You should add this as caption. Please let me know, because I would like to use it for an upcoming presentation, as it is quite a nice graph.

It is from my script "Einführung in das symbolische Rechnen". e.g.
http://www.informatik.uni-leipzig.de/~graebe/skripte/esr12.pdf p. 27

hgg

--

Dr. Hans-Gert Graebe, apl. Prof., Inst. Informatik, Univ. Leipzig