Manifold Systems
Graham Smith
incredible bargain. Has anyone got any views about this program good
and bad. On paper it competes well with Arcview/Mapinfo but in practice
is it as good as it looks.
Many thanks,
Graham
Agustin Estrada-Peña
I was a MapInfo 4.5 user and now I use Manifold for the same tasks, better
and faster in most of them. I have some cons, as it uses a lot of memory,
and lacks an e00 exporter (just some cons).
Moreover, some of the MapBasic freeley available programs currently
complementing the MapInfo functions are included as free solvers by the
Manifold's people. Just an example. The "conversion of many polylines to
regions" work I was performing with the help of MapLogix (sold separately).
To convert some 35,000 polylines to regions was a job of some *7 days* in my
Pentium II 400 Mhz, 128 RAM, but its is a matter of some *4 hours* with the
solver included in the Manifold 4.5 release.
The release 5 is coming, supporting raster functions, too.
Hope this helps.
Agustin
______________________________________
Prof. Agustin Estrada-Pena
Dept. of Animal Pathology (Parasitology)
University of Zaragoza (Spain)
Graham Smith <graham...@myotis.co.uk> escribió en el mensaje de noticias
VA.0000000...@myotis.co.uk...
Graham Smith
Agustin
Thank you for your views, they are useful. It does seem to be well
worthwhile pursuing and I have ordered a copy.
Cheers,
Graham S
Will O'Neil
seems like a strong, easy-to-use all-round GIS and a terrific value for
the money.
Will O'Neil
* Sent from RemarQ http://www.remarq.com The Internet's Discussion Network *
The fastest and easiest way to search and participate in Usenet - Free!
Graham Smith
Cheers,
Graham S
Nick Nicholas
Graham Smith <graham...@myotis.co.uk> wrote in message
news:VA.0000000...@myotis.co.uk...
> I am sure this has been aired here before, but manifold seems to be an
> incredible bargain. Has anyone got any views about this program good
> and bad. On paper it competes well with Arcview/Mapinfo but in practice
> is it as good as it looks.
>
> Many thanks,
>
> Graham
I've been using Manifold for some months now and find its pretty good
generally, Learning curve is huge though - there's just so many things it
can do. In capabilities it leaves both Arcview and Mapinfo for dead.
Nick Nicholas
Graham Smith
> I've been using Manifold for some months now and find its pretty good
> generally, Learning curve is huge though - there's just so many things it
> can do. In capabilities it leaves both Arcview and Mapinfo for dead.
Thanks Nick, I am now waiting impatiently for my copy to arrive.
Is your opinion on Arcview and Mapinfo based on experience of these two
programs or on a comparison of specs.
Cheers,
Graham S
Lorne Dmitruk
To be honest, comparing Manifold to ArcInfo is probably not fair. But if your going to be using large datasets I would use ArcInfo. One thing I would really like to see is a side by side comparison of Manifold, ArcView, and Mapinfo.
"Agustin Estrada-Peña" wrote:
Yes, Graham, sure it is.
I was a MapInfo 4.5 user and now I use Manifold for the same tasks, better
and faster in most of them. I have some cons, as it uses a lot of memory,
and lacks an e00 exporter (just some cons).
Moreover, some of the MapBasic freeley available programs currently
complementing the MapInfo functions are included as free solvers by the
Manifold's people. Just an example. The "conversion of many polylines to
regions" work I was performing with the help of MapLogix (sold separately).
To convert some 35,000 polylines to regions was a job of some *7 days* in my
Pentium II 400 Mhz, 128 RAM, but its is a matter of some *4 hours* with the
solver included in the Manifold 4.5 release.
The release 5 is coming, supporting raster functions, too.
Hope this helps.
Agustin
______________________________________
Prof. Agustin Estrada-Pena
Dept. of Animal Pathology (Parasitology)
University of Zaragoza (Spain)
Graham Smith <graham...@myotis.co.uk> escribió en el mensaje de noticias
VA.0000000...@myotis.co.uk...
> I am sure this has been aired here before, but manifold seems to be an
> incredible bargain. Has anyone got any views about this program good
> and bad. On paper it competes well with Arcview/Mapinfo but in practice
> is it as good as it looks.
>
> Many thanks,
>
> Graham
>
-- Lorne Dmitruk "Music is your own experience,your dmi...@worldgate.com thoughts,your wisdom. If you don't live it, it won't come out your horn." Charlie Parker
Dimitri Rotow
>Just ran a test this afternoon, using a "dinosaur", ArcInfo 8.0.1. I ran the
>test on a PII 266 with 64 MB ram, NT 4.0 sp4. Dataset 140220 arcs, to build
>polygon topology 3 minutes, to clean topology 5 minutes. Total number of
>polygons created 29138. A couple of years ago I created a dataset that
>consisted of about 1.5 million polygons on a Sparc20 with 64 MB ram, it took
>about 2 hours to create.
>
>To be honest, comparing Manifold to ArcInfo is probably not fair. But if your
>going to be using large datasets I would use ArcInfo. One thing I would really
>like to see is a side by side comparison of Manifold, ArcView, and Mapinfo.
>
Actually, it's most likely that ArcInfo was running this job slower
than Manifold would. Manifold is often faster in big jobs (when run
on NT, a faster system than '95 or '98) than legacy systems.
140,220 arcs sounds impressive, but it is likely orders of magnitude
less than 35,000 polylines. That's because *each* polyline is likely
equal to hundreds or thousands of arcs. An arc can be just two
inflection points, the beginning and the end. In contrast, a
polyline is most likely hundreds or thousands of inflection points,
depending on the complexity of the data set.
A consideration of possible algorithms shows that the time to build
polygons grows faster than a simple linear increase based on the
number of inflection points that define the polygon. This is true for
AI and Manifold and any other GIS as a consequence of elementary
geometry. Bottom line: it's fast to create polygons from
straight-line boundaries and entirely a different deal to do so from
curvy boundaries.
I don't know the exact nature of your data set, so I can't say whether
Manifold or AI would be faster in this particular case. However, the
previous posting did use a direct "apples to apples" comparison in
reporting that Manifold took hours for what used to take days. This
is a fairly typical result when comparing the faster Manifold
algorithms to those used in legacy GIS systems.
One must remember that the Manifold team arose from a project that
improved the speed of Intel's massively parallel supercomputers by
applying better algorithms to fundamental mathematics computation.
The project increased the speed of square roots by over one
hundred-fold and a few trig functions by a thousand-fold. It was such
a big success that the team organized under the name "The Center for
Digital Algorithms" to do more of the same.
The founding projects were an 800 function graph theory library and a
200 function computational geometry library. All this is fresh,
state-of-the-art mathematics that benefits from math advances in the
last several years as well as from original work by the Manifold team,
and it is this algorithmic foundation upon which Manifold is based. No
one has faster algorithms.
So, I do have to agree with you, it's not fair to compare Manifold to
ArcInfo.... AI costs forty times as much, yet it uses an archaic
interface, takes a year to learn to use, can't interface to most
modern software, employs ancient algorithms, uses the computing
equivalent of Sumerian as a scripting language and often ends up doing
less than Manifold. :-)
Cheers,
Dimitri
PS: I was going to write "Sanskrit" as the language in the last
paragraph above, and then I realized this would be a wrong analogy.
Echos of Sanskrit have survived in modern Indo-European languages
whereas AML is an utterly dead end, more like Sumerian. ... it's good
to see that AI8 has finally moved to VB.
vi...@ice.ru
>>Just ran a test this afternoon, using a "dinosaur", ArcInfo 8.0.1. I ran the
> 140,220 arcs sounds impressive, but it is likely orders of magnitude
> less than 35,000 polylines. That's because *each* polyline is likely
> equal to hundreds or thousands of arcs. An arc can be just two
Not in arc/info terminology. In Arc/Info terminology Arc is polyline.
--
--------------------------------------------------
Victor Wagner vi...@ice.ru
Programmer Office:7-(095)-203-50-60
Institute for Commerce Home: 7-(095)-135-46-61
Engineering http://www.ice.ru/~vitus
Vincent Heurteaux
Ok! Manifold seems to be a nice product, But is there a chance to
install it in a near future under Linux or *BSD, or should i keep winNt
on my workstation to use well interfaced GIS ?
<Troll-Mode OFF>
Does a GTK-GRASS developement start since it's released under GPL?
Thanks
vincent
Lorne Dmitruk
>So, I do have to agree with you, it's not fair to compare Manifold to
ArcInfo.... AI costs forty times as much, yet it uses an archaic
interface, takes a year to learn to use, can't interface to most
modern software, employs ancient algorithms, uses the computing
equivalent of Sumerian as a scripting language and often ends up doing
less than Manifold. :-)
Also in terms of interfacing to modern software last time I checked AI can be accessed using the following: VB, Visual C++, Java, C++, Motif, TCL/TK. I would also check out the new desktop tools in AI 8. They look up to date to me.
Cheers,Dimitri
PS: I was going to write "Sanskrit" as the language in the last
paragraph above, and then I realized this would be a wrong analogy.
Echos of Sanskrit have survived in modern Indo-European languages
whereas AML is an utterly dead end, more like Sumerian. ... it's good
to see that AI8 has finally moved to VB.
-- Lorne Dmitruk "Music is your own experience,your
Dimitri Rotow
>
>Not in arc/info terminology. In Arc/Info terminology Arc is polyline.
>
True. I assumed given what we know of AI performance that he was
using the term to mean a simple link in the graph theoretical sense...
that was wrong.
I should have stuck to the general observation that time required to
build polygons from polylines depends on the complexity of the
polylines. So, the comparison only makes sense when the same data set
is used in both cases.
Cheers,
Dimitri
Dimitri Rotow
<v.heu...@tour-du-valat.com> wrote:
><Troll-Mode ON>
>Ok! Manifold seems to be a nice product, But is there a chance to
>install it in a near future under Linux or *BSD, or should i keep winNt
>on my workstation to use well interfaced GIS ?
><Troll-Mode OFF>
>
No chance for a Linux version. Use www.dejanews.com to search for
"Linux" and "Manifold" and you'll find a rant on this posted
01/05/1999 that explains why.
>Does a GTK-GRASS developement start since it's released under GPL?
>
Ummm... not sure what you mean. Manifold is not under the GPL,
although lots of maps for Manifold have been released under GPL.
- Dimitri
ian gillespie
wrote:
>
>Warning to everyone.... *long* rant ahead....
not required, it is assumed in this NG/thread! ;-)))
Ian Gillespie
Geomatics Unit
Atmospheric Environment Branch - Ontario Region
Environment Canada
(opinions expressed are mine)