[Gmsh] comments about tutorial files
oshima at eng.niigata-u.ac.jp
Thu Jan 14 12:54:51 CET 2010
Dear Gmsh developers,
Thanks for hearing our comments. We see and really appreciate that the
tutorials (especially t2.geo) have improved and are a lot more
This time here are our comments about t4.geo and t5.geo. As our
intended audience are real novices the followings are what we do
really consider, but we admit they may be overkill for usual users. We
appreciate if the Gmsh developers take some of them into account.
* It might be helpful for users if there are brief notes about each
function (Hypot, Sqrt, T2, T3, StrCat, Today).
* Line 90-91 "Since this surface has a hole, its definition now
requires two lines loops": Perhaps a bit more explanation is needed
that, if we have two or more line loops specified in a Plane Surface
command, all line loops other than the first one define holes.
* Line 9: be -> by
* Lines 21-29: In reality, the Windows version of Gmsh 2.4.2 generated
1322 nodes in volumes and 11,665 tetrahedra with -clscale 1
specified, and about 300,000 nodes in volumes and about 2 million
tetrahedra with -clscale 0.2 (the numbers are important for the
users to feel confident that they are operating Gmsh correctly). It
might be further better if there is a note that the numbers can be
checked via Tools->Statistics.
* Line 128: It's obvious, but would be clearer and draws reader's
attention if thers's an annotation that Call is the command to call
* Line 136 "the format string should only contain valid floating point
format specifiers": It may be more helpful to non-programmer users
if there is an explanation that %g is the floating point specifier.
* Line 156: Like t4.geo, "the surface loops other than the first one
p.s. Just noticed one more thing: the bottom of the revised t2.geo
// To save all the tetrahedra discretizing the volumes 119 and 120
should be 129 and 130.
Takuya OSHIMA, Ph.D.
Faculty of Engineering, Niigata University
8050 Ikarashi-Ninocho, Nishi-ku, Niigata, 950-2181, JAPAN
From: Christophe Geuzaine <cgeuzaine at ulg.ac.be>
Subject: Re: [Gmsh] comments about tutorial files
Date: Mon, 11 Jan 2010 15:07:08 +0100
> On 21/12/09 12:00, Takuya OSHIMA wrote:
>> We are doing some detailed review of the tutorial files (t1.geo -
>> t9.geo that come with Gmsh 2.4.2) as a process of creating teaching
>> materials for our FE code that uses Gmsh as preprocessing front
>> end. In the process, at the stage of the review until t3.geo, we came
>> to have the following comments (we hope we are not nitpicking). What
>> do you think?
> Dear Takuya - many thanks for your insightful comments. We've tried to
> address them in the new version of the tutorial
> https://geuz.org/trac/gmsh/changeset/7024 (login: gmsh passwd: gmsh)
> Let us know if you have other ideas to improve the documentation: we
> would be very glad to use them.
> Christophe and David
>> Line 62 "specifying their orientation": it might be better if we have
>> an explanation that this applies only for lines and surfaces, and
>> have an example in one of the following physical entity definitions
>> so that one can see how to specify the orientation.
>> Overall, one may hardly be able to follow the tutorial since there is
>> little explanation about which entities Points 8, 12, 16, 7, Lines
>> 102-105, etc. are referring to (and which exact process generated
>> the entities). If the entity numbers are to be obtained by saving
>> unrolled geometry as explained at the bottom of t2.geo, the
>> explanation might be better to come at the top of the file and to be
>> clearly stated as such.
>> Lines 32-33 "a list": it might be better if it is clearly stated that
>> the list is referring to tmp (obvious for programmers, but not for
>> Line 40 "news": the special meaning of the variable may be worth
>> being mentioned.
>> Lines 76-77 "`Tools->Options->Save as defaults' button": there does
>> not seem to be such a button on the GUI?
>> Takuya OSHIMA, Ph.D.
>> Faculty of Engineering, Niigata University
>> 8050 Ikarashi-Ninocho, Nishi-ku, Niigata, 950-2181, JAPAN
>> gmsh mailing list
>> gmsh at geuz.org
> Prof. Christophe Geuzaine
> University of Liege, Electrical Engineering and Computer Science