Installing Manifold was pretty easy as I had .NET already installed. After running Manifold for the first time, Manifold let me know there was an update and I had to uninstall the version I just installed and then reinstall the update. I like it when Setup applications check before installing rather than this route. That said installation was quick so it wasn’t too much of a hassle.
Now the GUI. I’m a fish out of water here. To be fair Manifold acknowledges this:
A Note for ESRI or MapInfo Users
As mentioned in the For Experienced GIS Users topic in the introduction, Manifold’s user interface is not at all like those used in older GIS products. Instead, Manifold’s nomenclature and commands are based on a combination of several stylistic approaches:
· Microsoft methods and terminology wherever a Manifold capability matches something within the Microsoft spectrum of products. For example, Manifold uses “Tools - Options” because that’s where user preferences are kept within most Microsoft products.
· Microsoft Visual Studio, Visual C++ and Visual Basic user interface style for development and programming.
· Adobe PhotoShop concepts for photographic/artistic image editing.
· Manifold System logic and mathematics for GIS.
The manifold.net team is often asked why we did not simply clone ESRI methods and interfaces in GIS, since ESRI dominates GIS markets. There are three reasons why:
· The potential user base for GIS is far larger than the current ESRI population, which is a very small number by mass-market Microsoft standards. Making Manifold accessible to hundreds of millions of Microsoft Office users means adopting Microsoft terminology and methods, not ESRI’s.
· Older GIS systems were created in a technologically less sophisticated time. Providing the full range of modern capabilities means adopting modern ways.
· We seriously intend to deliver a fusion of GIS, CAD, advanced database, image processing, photo editing, mathematics modeling and many other disciplines. The right user interface that works with a blend of such intense functions is a new user interface that blends the best ideas from all these disciplines and not just GIS.
We realize that the above approach make transitioning to Manifold more difficult for experienced GIS people who have a lot of ESRI or MapInfo experience they would like to leverage. However, for many ESRI or MapInfo users a transition to Manifold is really their first transition into the bigger world of Microsoft standards and mass-market PC software standards. For any professional person using computers this is a necessary step and not wasted effort.
I’m not going to post too much on Manifold and how it compares to GIS systems that I’m used to until I get some practice with the application. I’m used to the ESRI way, good or bad, so I just need to spend some time and work though the new concepts. Smartly Manifold has a whole help section devoted to the “Experienced GIS User” which goes over the differences between the two products. I’ll find some time tomorrow to really get into Manifold before trying some more complex tasks.
DISCLOSURE - This copy of Manifold was provided to me by Manifold for evaluation.
31 responses so far ↓
1
Kirk
// Dec 13, 2006 at 2:45 pm
Hey James,
I’m anxious to see how this compares with ArcGIS. Are you really running on a Pentium III ?
Regards,
Kirk
2
Doug
// Dec 13, 2006 at 2:59 pm
“We realize that the above approach make transitioning to Manifold more difficult for experienced GIS people who have a lot of ESRI or MapInfo experience they would like to leverage. However, for many ESRI or MapInfo users a transition to Manifold is really their first transition into the bigger world of Microsoft standards and mass-market PC software standards. For any professional person using computers this is a necessary step and not wasted effort.”
I wonder when manifold getting the latest Office GUI look and feel, and if that would be a good or a bad thing?
To be honest it is hard not to roll my eyes at a GIS that thinks Gaussian blur is a useful feature. Yes that is what I do to my maps all the time… blur them.
3
Jacob
// Dec 13, 2006 at 3:22 pm
Are you kidding me?
http://www.manifold.net/doc/7x/terminology_in_gis.htm
Why does the name of a compeditor have anything to do with why manifold calls their polygons, areas?
Of course ArcInfo was around decades before Manifold so almost any term could be call “Archaic”.
4
James Fee
// Dec 13, 2006 at 3:28 pm
I’ll be honest, the documentation turns me off. Half of it is anti-establishment, but I guess that is their market.
As I said above, it is a totally different environment than the ESRI/Autodesk/Adobe world I’m used to. There is logic to the madness, and I’m begining to see it. Where as any ESRI user can pick up QGIS and run with it, you really need to focus on the documentation with Manifold. We’ll see what the payoff for reading the documenation is later this week.
5
Doug
// Dec 13, 2006 at 3:36 pm
Well Jacob you have found an excellent example of Manifold pulling some BS out of the air.
ArcGIS uses the word “edge” to describe graph theory edges. For example if you look at the geodatabase object model diagram you will see that “EdgeFeature” is a type of network feature, i.e. that has a from junction and a to junction. So ESRI uses “edge” according to the graph theory definition in the two places where their software deals with graphs: the geometric network and the network dataset.
As for calling a curvy line an “arc”, webster has several geometry related definitions. Two of them are “a: something arched or curved b : a curved path” so calling the line segments between line vertexes a path is fine.
Not all lines are used in a graph. Trying to use all your lines in a graph is a good way to get gobbledy gook. You would want to use your roads in a graph.
And if manifold is so interested in graph theory then maybe they should implement turns so that you can model the real world. Maybe they could implement a graph with a connectivity model while they are at it too. How about some multimodal algorithms plz? I guess all the “professional mathematicians working in graph theory” who use manifold are happy with road networks that don’t reproduce the real world.
6
James Fee
// Dec 13, 2006 at 3:37 pm
Kirk, no on a Dell Precision M70. I believe it has a Pentium M in it and something like 2.26 mhz.
7
Derek
// Dec 13, 2006 at 3:49 pm
Hello everyone. I’ve done a bit of reading on Manifold, but I have some questions. Does it have ESRI capabilities, such as, 3D and Spatial Analyst? When exporting to Shape or KML, does it work well? The web site talks about how Manifold is the next greatest thing in GIS since “sliced bread.” I do, however, love the price. I just wish they had a trial version I could play around with. Thanks!
8
Doug
// Dec 13, 2006 at 4:50 pm
Man it gets better:
“A network is defined by its nodes and not by lines. The most common error of graph theory newbies is to draw a system of lines that have coincident ends and then say, “there, that’s my network.” Nope. If it does not have any points it is not a network. One point by itself can be a network. Two points by themselves is a disconnected network. Two points with a link between them is a connected network. Any number of lines without points is not a network even if they are arranged in a pretty diagram. Even within archaic graph theory nomenclature a line is not an arc unless it has a node at each end.”
Pity the poor users who think that their street edges would be an input for a graph.
According to Manifold you are conceptually in error if you think that your roads (which are “edges”) are what are important in your graph. Further manifold perpetuates this silly bias by not leting you locate the inputs to graph problems (e.g. create drive time zones) along the edge. Your graph analysis points must be snapped to the edge endpoint to be usable. Manifold seems to be blithely indifferent to the fact that a graph is a collection of vertexes and edges and you should be able to use both in defining the graph and analysis.
http://www.manifold.net/doc/7x/drive_time_zones.htm
“Very Important: Points to be used as centers must be exactly coincident with the end of one of the lines in the road network”
So if you want to sart your route in the middle of a mile long country road you are out of luck as you’ll have to start it at an edge endpoint.
If you want to do a drive time analysis for your fire station in manifold (yay convex hull) your fire station had better be located at an edge endpoint.
Geocoded points of interest located along the edge? Hey you’re out of luck when it comes to using Manifold.
“Since the vast majority of GIS data that uses line objects consists of lines drawn for purely illustrative purposes without points it is mathematically illiterate to refer to such objects in a general way as arcs. Even those few GIS data sets in which lines have points at the ends of them are not often graphs in a consistent way.”
I’m sure that NavTeq and TeleAtlas would laugh if they were told that their road data wasn’t a properly consistent input for a graph. Dimitri is a smart comrade, he can probably work out how to infer endpoints in a planar graph and indeed he does take planar lines as an input to a graph operation:
“The drawing of road lines used must be a true road network that is a connected network. If there are small breaks between the road lines or if the ends of adjacent lines are not exactly coincident the drawing cannot be used as a road network”
So in the end, Manifold is a real GIS for graph gearheads that uses vertexes to define the network, not lines. And then when you have to supply a network you point Manifold at some lines with coincident endpoints which will be used to make the graph. Which way is it Dimitri? Your doc says “A network is defined by its nodes and not by lines” and yet ask the user to provide “a set of lines to be used as a road network”. Maybe you should cut the holier than thou stuff in your documentation and just document your system.
Oh and calling your prospective users “newbies” is so late 90s. The proper term would be “noobs”.
9
Mike
// Dec 13, 2006 at 4:55 pm
I found this a while back and added it to my collection. http://dspace.library.cornell.edu/retrieve/562/review_final.pdf
It is a pdf that compares common tasks between arcgis and manifold.
10
Rollo
// Dec 13, 2006 at 4:57 pm
I’m not an Esriite, but I also find Manifold impossible to use. All the functionality is hidden away in obscure contextual popup menus; functions with frankly bizarre names; and toolbars with odd symbols.
When you open up Manifold, you are confronted with an array of toolbars - lots of tools which clearly have similar functionality; and many which seem bizarre altogether. Importing and viewing a shapefile (for instance) is then a counter-intuitive series of five or six commands, hidden away in different contextual menus. I’m afraid that while I have other GISes installed on my desktop, that alone is enough to put me off regular use of Manifold.
And it’s not for want of trying - I really wanted to love Manifold; just as I really want to love Cadcorp. The difference is that Cadcorp actually listen to what their users want, and provide accordingly. Shame about the lacking functionality though…
I have yet to see a perfect GIS. There are the overpriced, stuck-in-the-dark-ages; globalising dinosaurs; or the up-and-coming-but-lacking-functionality newbies; or the cheap-and-nasty, over-defensive rebels. Take your pick. There’s nothing wrong with shopping at Aldi or Lidl just because they’re cheap!
I can understand why there’s a need for a budget GIS, and why people might use it, but they seem to try to alienate everybody at every possible opportunity - the quote in your post is a fine example of this - if they know people are finding their software difficult to use, they should do something about it, not justify the way it looks/runs.
11
Chris C.
// Dec 13, 2006 at 6:18 pm
‘Importing and viewing a shapefile (for instance) is then a counter-intuitive series of five or six commands, hidden away in different contextual menus.’
Let’s see:
1. File> Drawing Import, select SHP file, hit OK.
2. Open drawing component.
3. Open table component.
Done in three steps, unless there’s no PRJ file for the archaic format and you need to use Assign Projection, which would make four.
‘functions with frankly bizarre names’
If you don’t understand what all the various commands do that’s where Help comes in. Manifold isn’t an ESRI clone, it has its own terminology, like it or not.
‘And it’s not for want of trying - I really wanted to love Manifold; just as I really want to love Cadcorp. The difference is that Cadcorp actually listen to what their users want, and provide accordingly. ‘
Manifold has a suggestions topic in Help on how to provide user input on possible improvements and additional features. Over on the Georeference forum you’ll find threads which confirm users wishes are being heard and acted upon.
‘I also find Manifold impossible to use’
If that’s the case then maybe Manifold isn’t for you, but there many users out there who find it easy to use.
12
barb
// Dec 14, 2006 at 12:30 am
Another happy Manifold user here (and a graph junkie, too).
Doug, you appear to be taking things out of context. The help topic talking about nodes that you cite contains this phrase at the very top: “A network consists of nodes that may or may not be connected together by links.” There obviously is room for improvement in every help topic, but yours saying that “according to Manifold you are conceptually in error if you think that your roads … are what are important in your graph” is totally out of the line. The help topic fully acknowledges the fact that links are important. All it is trying to do is make you aware that a graph consists of something more than links and that having a pretty collection of lines is usually not equal to having a graph, in that the collection of lines will frequently not work well as a graph, and that in particular you should care about whether or not these lines are connected to each other. One of the easiest ways to visualise the connections is to have a physical point object at every location where one would want to see a node and so the topic advocates this. That’s good, because disconnections in what looks like a connected set of lines is I think number one problem that people run into when they start using graph theory algorithms on road networks!
As for drive-time zones, you absolutely can start in the middle of a mile long country road. Just be sure to use the “Split roads at intersections” option (on by default).
13
Doug
// Dec 14, 2006 at 8:03 am
Wouldn’t the “split roads at intersections” option also split overpasses and underpasses so that roads that cross are considered to be connected?
14
Jack Dangermond
// Dec 14, 2006 at 8:10 am
For all those Manifold haters, look at Orbit GIS…www.orbitgis.com. Under $400 and good functionality.
15
rmcculley
// Dec 14, 2006 at 8:18 am
If I don’t have an intersection in the middle of my mile long country road, how is “split roads at intersections” going to help?
16
Chris C.
// Dec 14, 2006 at 9:39 am
‘For all those Manifold haters, look at Orbit GIS’
I seem to recall you need to get add-ons for IMS capability, again you can’t compare apples to apples.
I’ll stick with $100 run-time of Manifold to serve my IMS sites.
17
barb
// Dec 14, 2006 at 10:46 am
The “Split roads at intersections” option will split roads at both intersections and points selected as “centers”.
Yes, using the option will split roads at all intersections. If you want to split at some intersections, but not others, you have to split manually and turn the option off. Sounds logical to me.
18
barb
// Dec 14, 2006 at 10:48 am
You know, perhaps James could compare Manifold to Orbit GIS for us. What do you guys think? No tongue in cheek at all, by the way. If there is something that Orbit GIS has that Manifold does not have, I will absolutely suggest it to the Manifold team. They are a responsive bunch.
19
Doug
// Dec 14, 2006 at 10:52 am
So if I want to stop/start/ a route along an edge I have to split intersesctions, but if I do this then I would by default negate the differences between overpasses and underpasses (i.e. interpret my non planar data as planar data)? Talk about Faustian functionality…
I suppose you could manually split the edge that has a geocoded destination, and then solve without splitting intersections, but that seems like a good deal of work (and editing) to get a route from A to B.
20
barb
// Dec 14, 2006 at 11:02 pm
Not quite.
Suppose you want to preserve overpasses and underpasses and thus can not use the automatic split option.
If you want to start in the middle of an existing road, you have to split that road. This is very simple to do using visual tools, but if you have to do this often enough, you can automate this to a one-click action using a script.
Ending in the middle of a road is always permitted no matter what, and thus is not an issue.
I agree it might be useful to have one more option that would only split roads at the starting location and nowhere else, however given how easy you can get this with the existing functionality this is a minor point for me.
21
barb
// Dec 14, 2006 at 11:06 pm
Um, and if you want to get a route from A to B, would not you already have the roads split at A and B?
22
Dirk
// Dec 15, 2006 at 9:02 am
Dudes! Can anyone actually tell me anything useful on Manifold?
23
barb
// Dec 15, 2006 at 9:22 am
This sounds like a troll, but what the heck…
Manifold is a GIS that can handle vector and raster data. It contains spatial SQL, integrated IMS, an elaborate object model for script writers, 3D terrains, analysis tools, publishing tools and a wealth of other things. It can import and export data in a variety of formats and bring data from external data sources via ODBC, OLE DB and other database technologies. It can run natively on 64-bit machines, too.
There are literally thousands of features, some of them seen in other GIS products and others unique to Manifold. The development cycle is pretty fast with the company incorporating hundreds and thousands of features suggested by users every year (simplifying things here). The number one Manifold location on the net (after http://www.manifold.net) is http://forum.manifold.net.
Was that useful?
24
KoS
// Dec 15, 2006 at 10:02 am
I don’t know why I thought of this. But after seeing Dirk’s post with dude in it.
Dude I got Manifold……not to rip off Dell.
KoS
25
Doug
// Dec 15, 2006 at 10:10 am
“The “Split roads at intersections” option will split roads at both intersections and points selected as “centers”.”
So this option means that you can’t get any kind of accurate routes from your data. You would jump off of overrpasses onto underpasses. Your routes would not be accurate, would be unusable and would be dangerous to follow in the real world.
“Yes, using the option will split roads at all intersections. If you want to split at some intersections, but not others, you have to split manually and turn the option off. Sounds logical to me. “
Why do I have to edit my roads to perform an analysis on them? This is a key sign of a poorly thought out system- that it requires a user to frequently edit their data to perform an analysis. In this case, you would have to manually split everry edge that you want to route from, properly apportinoing the attributes. Do you see how unworkable, and frankly, how much of a hassle/gimmic this scenario becomes? Just because a story is “logical” doesn’t mean that it is a valid workflow or that it should be that way. Even so, why is it “logical” to have to split an edge so that you can start a route along the edge? ESRI software doesn’t require this step. If this is the kind of consistency Manifold offers then i’ll need to lower my estimate of it as a product.
26
Doug
// Dec 15, 2006 at 10:19 am
“Um, and if you want to get a route from A to B, would not you already have the roads split at A and B?”
No. By extension, given a route between two arbitrary addresses (think google maps), then I would have all my edges split at all points along them so that A and B could snap to a junction, making for nearly infinite edges and junctions.
Most things in the real world happen on the edges (that is where packages are delivered to, where people live, etc).
Imagine not having to have A or B located on a junction. Imagine just being able to place them and have them snap to the nearest edge or junction (whichever you choose). Imagine how much time you would save without splitting. Imagine how few changes you would make to your data when you are routing. I think the real kicker is that there is no technical reason why you can’t locate A and B on an edge aside from the fact that the manifold programmers are lazy and didn’t want to modify the Dijkstra algorithm to handle edges (it natively is a junciton-junction algorithm).
27
barb
// Dec 15, 2006 at 11:04 pm
Please, Doug.
If you want to be able to turn at any intersection, use the split option. If you do not want that, do not use the option. It is that simple. There is absolutely no need to moan that using the option “means that you can’t get any kind of accurate routes from your data, you would jump off of overrpasses onto underpasses”. The option does not “mean” that. If you have overpasses and underpasses you should not use it.
Let’s turn to starting in the middle of a road. As I already said, this is a simple edit of one line, easily automated using a script. As I already said, I understand that you might not feel comfortable making this small edit and I agree that it would make sense to have an option to split roads at the starting location and nowhere else. Nothing new here. A small problem with an easy fix and a small feature request.
Now, is the requirement to start in the end point of a road (or, equivalently, the requirement to manually split the road containing the starting location prior to doing the analysis) - in case you want to preserve overpasses and underpasses - the only gripe that you have, or do you have something else? You make it sound so big.
28
Dirk
// Dec 18, 2006 at 7:15 am
Dude!
29
Dirk
// Dec 18, 2006 at 7:18 am
Dude, don’t mess with the Dangermond, dude.
30
Fordprefect
// Mar 28, 2007 at 5:36 am
Doug said:
“To be honest it is hard not to roll my eyes at a GIS that thinks Gaussian blur is a useful feature. Yes that is what I do to my maps all the time… blur them.”
You show your ignorance . Gasussian Blur operates on images, not vector drawings.
You clearly know next to nothing about image processing. Gaussian blur used with other processing procedures allows reduction in intensity variations across an image. It can also be used as an artistic special effect or when used with gridded data, to create a smooth transition between isolated data values.
31
ricky
// Mar 29, 2007 at 6:27 am
Ford,
you should not be surprised by the comment. This is an ESRI blog. Therefore, the people on it know very little. They are indoctrinated into ESRI-speak. If ESRI can’t do it, then it isnt important, right?
I’ve not used Manifold myself, but I get the sense that Manifold users are alittle more computer saavy than ESRI users. ESRI has lowered the bar so that all of our expectations are in the basement. So, the tools that Manifold offers, if they actually work, are ignored by the ESRI community because they are ignorant of what advanced GIS could be.
Leave a Comment