Question from my father who is looking to explore alternatives to Manifold:

"I am an independent environment consultant who is a heavy user of GIS for spatial data crunching, geo-referencing self-digitised maps, terrain modelling using LIDAR and analysing GPS field data. I have a workstation 8 core Intel i7 3.4 GHz, 16 Gb of RAM, NVIDIA GeForce GTX 570 running 64 bit Windows 7 ultimate and dedicated HDs for system and cache files. For many years Manifold GIS has been my software of choice due to intuitive GUI, wide range of functions, programmability, natively 64 bit and low price. However, the current version (Manifold 8) is now 4 years old and is now very slow for many tasks. See here for a recent vector processing comparison where Manifold comes out 40x slower than ArcGIS. While Manifold has an impressive suite of surface rendering and analysis functions it is so slow in processing LIDAR data that is almost unusable for anything other than very small data sets. Manifold is CUDA enabled for a limited set of terrain processing functions (see the manual > “index” > “N” > “NVIDIA”) but not the actual rendering. One reason why the next version Manifold 9 is taking so much time may be because the big chunks of software are being coded for CUDA which everyone acknowledges is difficult. Manifold’s loyal user base has been promised the massively faster and improved Manifold 9, but there has been no sign of this new release in 2 years. I, like many Manifold users, are beginning to despair of any development of the software. To protect my business and keep ahead of the game I am being forced to look for other GIS solutions.

An alternative to Manifold would have to have the following:

  • Run natively in 64 bit and so address large amounts of RAM (ie < 4Gb and typically 16Gb and above).
  • Able to use the GPU for image/surface intensive processing eg CUDA on Nvidia graphics cards.

  • Terrain modelling and LIDAR processing capabilities, be able to ‘drape’ vector and raster map components and do 'fly throughs’ for video presentations.

  • Be a stable release.

Manifold 8 professional with ‘surface extensions’ costs $590 (£370). None of the options that I have seen so far seem up to the job. These include:

  • ArcGIS – I have just been quoted £1495 + £425 per annum to access upgrades and another £2,495 for the 3D Analyst extension.
  • QGIS – Seems only to support 32 bit in the stable 1.8.0 Windows version, buggy, and lacking in-house digital terrain mapping features.
  • R's spatial packages – this option seems to be too fiddly and static for what I want to do.

Looking forward to responses from people who have faced similar dilemmas – what's the best solution out of these options or should I go for something else completely?"

  • for the specific requirements and use case which is mentioned I would be expecting to pay thousands of $$. – Willy Dec 3 '12 at 6:46
  • Virtual Machine Linux or Windows. Depends how much Windows software actually use 3D graphics. If not much then windows as guest, if alot then linux as quest. I would also add FME to toolbox (about 2k€ + yearly upgrades for x€). – simplexio Dec 3 '12 at 8:10

The full version of ArcGIS (ArcInfo) with all the extensions is far beyond anything that Manifold can offer, but it is also more than ten times as expensive - so that is a fair comment. The base-price listed above even seems low for the full version so I suspect it is for an ArcView or maybe ArcEditor licence rather than ArcInfo.

QGIS does have digital terrain mapping features, especially when used in conjunction with GRASS (which is now built-in). QGIS is also not especially buggy so that comment is unfair.

GRASS has the ability to do flythroughs and has an impressive array of raster processing tools plus avast number of plugins that make it rival ArcInfo with Spatial Analyst and 3D Analyst plugins combined. GRASS, as a stand-alone GIS rather than a plugin to QGIS, has a 64-bit version for Windows. GRASS has also been around for decades and is very stable.

I do a lot of 3D work and even when I used ArcInfo, I didn't use it for flythroughs because it is too limited for a decent model. I build my own models using Blender and GDAL. Anybody working in the GIS field who is doing more than basic mapping really should not expect to have a one-size-fits-all solution.

As a freelancer I have a toolkit that contains at its core (in no special order):

  • QGIS (with GRASS)
  • Python with NumPy and SciPy
  • Mapnik
  • PostGIS 2 with PgRouting
  • SpatialLite
  • OpenOffice (you have to write reports to go along with the maps)
  • Blender
  • Unity
  • GIMP 2

I also use a few other more specific packages relating to windfarm work for instance. My point is that you can certainly achieve everything you can in ArcInfo (and therefore Manifold) using FOSS4G software, but you have to be more resourceful and technically adept than if you rely on one of the proprietary products and that is what you are paying for.

EDIT: Just for comparison, as a salaried GIS Professional I was used to having a core toolkit as follows:

  • ArcInfo with Spatial Analyst and 3D Analyst
  • ERDAS Imagine (for 3D stuff it's better than ArcScene)
  • Python
  • Photoshop
  • MS Office

I can do everything with my freelance kit that I could with the list above... and I'll leave it for you to do the price comparison :)

  • 2
    Have you tried GRASS 64 bit though the Sextante QGIS plugin (or built-in plugin if you are using > 1.8)? – Nathan W Dec 3 '12 at 10:36
  • 1
    Many thanks for this detailed and insightful answer. I have long been encouraging my dad to 'go open source'. I think he's a lot more likely to listen to a independent GIS professional than his idealistic son. The biggest barrier I suspect for him and others is relearning a totally new set of skills. In my limited experience (Windows to Linux, ArcGIS to QGIS) I have found this relearning experience can be very rewarding, if you have the time to invest. – RobinLovelace Dec 3 '12 at 11:15
  • I'm running all the above on Windows, which cuts down one of the barriers to going down the FOSS route. For me, Python is the Glue that holds it all together. I can get all these bits and bobs talking to each other with their respective Python APIs. This makes it all more cohesive than first appears. – MappaGnosis Dec 3 '12 at 13:19
  • 3
    I would add that the freeware, SAGA GIS has very robust surface modeling options and have 64-bit compiles for windows and Linux. There are also options for working with LAS data. – Jeffrey Evans Dec 3 '12 at 16:47
  • Yes, SAGA looks good, though I haven't used it enough to comment. There are also many other good bits of FOSS4G kit out there so my list is far from exhaustive - just what I have personally found to offer a good alternative to the proprietary kit I used to use, – MappaGnosis Dec 3 '12 at 17:43

While Manifold gives you a complete package, QGIS has many of the capabilities you would look for. There is the basic raster and vector handling, very easy integration with PostGRES, and even spatial SQL (but for that, you will have to have the data stored in PostGRES).

Spatialite is a minimalistic tool that has some really great spatial SQL capabilities. However, spatialite requires you to explicitly reference to spatial indexes to get any decent speed performances.

With python, you can also program in QGIS. There is a very granular object model. The documentation is such that you'll have to understand how OO programming to effectively navigate through it.

Geoserver, working with OpenLayers is a very nice alternative to Manifold IMS. But, I really like Manifold IMS, so most other things are pale in comparison.

Given that Manifold has been without a release in 4 years, and becoming dated in its capabilities, I am doing a lot more with QGIS and these other tools.

  • 1
    Same here, and QGIS has 64bit versions available for all platforms now as well – chrismarx Apr 7 '14 at 15:08

I think the only answer that can work is to do what every other low-cost user does, expand your toolkit with a suite of tools and learn new skills. There are plenty of lists of alternatives.

I would say that the criticism of QGIS being buggy is unfair, certainly here it is just anecdotal and so inappropriate here.

Not the answer you're looking for? Browse other questions tagged or ask your own question.