Subscribe to this thread
Home - General / All posts - Radian Studio Beta / Release 9 Outlook
adamw

7,031 post(s)
#24-Dec-15 12:34

This is a discussion thread for the announcement of Radian Studio Beta and general outlook for Release 9.

The beta is starting in the beginning of January.

See you on the other side. :-)

Donmap
205 post(s)
#25-Dec-15 05:35

Great News for the upcoming year 2016

DomagojC6 post(s)
#16-Feb-16 14:29

Hi!

I am very interesting about new software.

Please let me know how can be one of beta tester?

adamw

7,031 post(s)
#17-Feb-16 12:50

The application process is explained in the announcement.

dmbrubac


1,602 post(s)
#25-Dec-15 14:48

This is great news Adam!

I'm looking at various platforms to process point cloud data and I'm wondering if Studio is appropriate. My clouds range in size from 100 to 500 million points. They are dense matched points from UAV imagery as opposed to lidar, so everything is single return, but noisy. Intelligent cloud decimation through precise classification and processing is the order of the day. Right now CloudCompare is my tool of choice but I haven't written any plugins or automation scripts. Alternatively LASTools looks promising but comparatively expensive. There is a divide between products that put all the points in a bucket and let you work with them spatially or with attributes and tools that try to add intelligence. I'm good with the former.

I guess I have two questions: will there be or is there support for LAS / LAZ (easy to add if not - see http://rapidlasso.com/lastools/ and the open source LASZip project,) and do you think Studio would work? If it does, the timing is ideal. I'm looking to overhaul my processing pipeline this winter if the weather ever gets bad.


Don't expect, suggest!

adamw

7,031 post(s)
#25-Dec-15 15:00

I'd say it is worth a try - ie, 500 million points isn't really scary. Obviously, there's a lot to gain from organizing the data of that size in a way that uses least space and is best for the analysis you want to perform, but we are talking about faster or slower processing speed, not about painfully trying to fit into some hard limit in order to just load / work.

I don't think the file format is going to be a big issue - we might already support it, but even if we don't, it might be easy to add, as you say, plus I see 'lasgrid' in the link you provided.

Good to hear from you! Happy Holidays!

dmbrubac


1,602 post(s)
#25-Dec-15 15:53

Thanks Adam! Happy Holidays to you and yours!

Careful with the info in the link. Some tools are free and open source, others will introduce errors in the output at some millions of points processed without a license. LASindex and LASsort are important when using LASTools too. BTW CloudCompare builds an Octree index.

By the way, ESRI has a zLAS format that is nearly identical to LAZ but it's closed... Not that I expect Manifold to go that way, but it could be important info.

I think 'it's worth a try' is exactly the level of feedback I was hoping for. My highest priority has to be my processing pipeline and if I can do that and also participate fully in the beta then that a win for everyone.


Don't expect, suggest!

RonHendrickson
241 post(s)
online
#25-Dec-15 22:05

You Manifold Elves have been busy....and rascally in breaking the news on Christmas Eve.

Delightful! I can't wait to get into the new technology and see what is possible.

dmbrubac


1,602 post(s)
#26-Dec-15 14:05

Here is a link to a thread at the CloudCompare forum that talks about extracting bare earth (DTM) from a cloud of points representing a surface: http://www.danielgm.net/cc/forum/viewtopic.php?f=14&t=679&sid=c3bc3c9fb2f4f9d337123cca6253fd6e

I think replicating the 'Octree level selection' is a fairly simple SQL construct (define a cubic step size, or a square step size and a step on an attribute e.g. Z) and then taking the min, max, average, count, etc. Is even easier. If you read to the end, min and max functions will return a point either in the centre of the cell or at an existing point. I want to extend the average functionality to return an existing XY nearest the centre of the cell with an average Z. The solution in CloudCompare is to 'hack the code' instead of write a bit more SQL. I think you know what I'd rather do.

Man, I can't wait to try this!


Don't expect, suggest!

adamw

7,031 post(s)
#27-Dec-15 08:00

Exactly, it's pretty useful to be able to try something fast with SQL to see if it works or not instead of spending much longer writing low-level code to do the same.

I read some more about LAS/LAZ, I will say several things:

(a) we don't currently support the format so at first we will likely be talking about consuming exports into other formats, raster or vector - we do have the format on the wishlist though and even have some preliminary work done for it,

(b) it's not a big deal to finish the preliminary work that we have and support the format - if that happens, we'll support the format natively (and if it doesn't happen or doesn't happen fast enough, there's another way),

(c) a lot depends on what you want out of the cloud - the words 'decimation' and 'classification' do provide a general direction, but there's a lot of flexibility in possible approaches, with different approaches providing different benefits - it might be useful to take a look at the whole of what you are after, from start to finish.

I suppose we'll be able to talk about this more in the beta. :-)

dmbrubac


1,602 post(s)
#27-Dec-15 16:45

Yeah I think we can have a thread in the beta for sure, since I have links to pages that describe formulae for many of the different things I want to do, but I think it would only be the two of us talking.

By the way there is a google group for LASTools and recently someone posted a link to a VS2010 LASlib dll solution (https://drive.google.com/file/d/0B7HD2rT6Exl1SmFXbko1VzNDN1k/view) so even if you don't support LAS/LAZ natively, I could link this dll into a project. All I need to do is create a set of tables for the data.

Oh what the heck - someone else might be interested.

Hidden point removal

Classification

Planar facet extraction

Most of the other stuff I want to do is in the PCLlibrary

I will post a step-by-step description of what I do / want to do in the beta, but a somewhat advanced requirement would be to extract an accurate top of rail, since I'm doing some work with rail reconstruction and expansion.

Basically I trace the rail on the ortho I get out of the drone, then use the 2D polyline to select points out of the cloud. Then I use the points most likely to be on the top of the rail to define a pair of 3D polylines that best fit tangents or arcs in 2D, and regular slopes in 3D. I also must ensure the gauge of the railway stays at the standard distance. On tangents I want to make sure the Z value at any given section line is very close, but on curves, where you can have super-elevation, I don't worry about calculating whether or not the super-elevation makes sense. On main lines I set my tolerances very tight but on sidings and spurs things get a bit looser.

Since I can easily have a 100,000 points that represent each rail over a couple km stretch, I end up with a result that gives me a statistically meaningful representation of the rail in 3D.

This is pretty manual right now...


Don't expect, suggest!

tjhb

7,283 post(s)
#27-Dec-15 18:20

someone else might be interested

Very.

artlembo


2,784 post(s)
#27-Dec-15 20:04

but I think it would only be the two of us talking.

We actually teach a course using laztools. I don't teach the course but one of my colleagues does.

The course and the results of that have gone hundred enough interest from the state of Maryland that we will probably be doing a lot of lidar work with them.

On another note, we manage the lidar server for the entire state using Arcserver. So I have all the statewide TMs ready to go to see how well radian can serve them up.

dmbrubac


1,602 post(s)
#27-Dec-15 21:08

Oh well in that case! Hopefully 3 Manifold 'Lifers' can sway the powers that be into native support. *Edit 4 now - hi Mike!

I think LASTools is a really good piece of software but I don't want to spend 5,000 Euros and a cumulative 20% per year (1000 in year 2, 1200 in year 3, 1440 in year 4 etc) in maintenance. I also think it's far better suited to a different kind of project than my clients want.

Radian could be revolutionary to point cloud processing because, honestly, the fun and easy part of my job is flying the drone. The fun and hard part is turning that mountain of data into outstanding products. If I can perform SQL queries against my cloud to achieve what I need then I couldn't be happier.

Adam, here is a simpler use case that is actually quite complex in its own right.

The file RGP Cloud clip is screen shot of a section of a cloud containing about 7 million points. It is a freshly paved parking lot enclosed by a fence and surrounded by flood lights. The parking lot contains newly built truck bodies, so the data is pretty noisy. Also, there are voids in the data because of the characteristics of the asphalt. You can also see voids in the deep shadows around the truck cabs.

The Scalar Z Cloud Clip file shows the same data coloured by elevation. You can see the heads of the lights, some of the posts are picked up as are sections of the fence.

Lastly, the Scalar Z Cloud Detail file shows a small part of the point cloud (400,000 points) with the normals computed. You can see how 'soft' the trucks look and a ridge where the shadow of the light falls across the parking lot. The 'flat' parking lot... isn't.

There is, however, a trend to a lower elevation to the upper left. There is a catch basin up there somewhere. I want to take out the trucks and smooth the parking lot so that it represents reality significantly better than what you see in the clips.

Attachments:
RGP Cloud Clip.JPG
Scalar Z Cloud Clip.JPG
Scalar Z Cloud Detail.JPG


Don't expect, suggest!

jkelly


1,234 post(s)
#04-Jan-16 02:36

I'll add my weight in here as well, as we got ~20000Ha of LiDAR flown in the middle of last year. At this stage, we are only utilizing the derived products (DEM, DTM, DSM, Hillshade etc) and have not delved into the raw point clouds, but it would be great to have some other options for crunching the data.


James Kelly

http://www.locationsolve.com

mdsumner


4,178 post(s)
#27-Dec-15 20:57

Just to throw in the mix:

PDAL: http://www.pdal.io/

Format contentions: http://rapidlasso.com/2015/12/19/the-darc-force-awakens-esri-escalates-lidar-format-war/

From my perspective, Manifold needs a huge re-working of the data representation to deal with "point clouds", or any serious 3D (or higher) data sets. I'd hate to see a new kind of layer for extending beyond X, Y - which allowed "Z", and RGB, 'gpstime' and so on for point clouds only, while still treating X,Y as somehow privileged. Where would non-numeric point-attributes be put? If Manifold gets a "multi-point" type it makes it clear that vertices hold these attributes, not the objects.

Anyone else have strong opinions here? (I don't see any other package that approaches this generally enough, though the PostGIS structures are the closest).

All this fits much better into a flexible relational-table structure than any especially spatial data structures that are out there. Spatial object hierarchies are far better expressed in tables and links than in nested/recursive structures IMO since it's so much more customizable.

I'd love to see a generalization of the Drawing that can deal with arbitrary sets of attributes (beyond X, Y) for vertices, but it also needs a generalization from from branch-based polygons, lines and points to "surfaces" (not grids or TINS), lines and multi-points that can be seen as sets of branches OR as collections of primitives. This is where the cross over lives for visualization as well, the graphics needs the primitives.

Hoping this is already in the works. :)


https://github.com/mdsumner

mdsumner


4,178 post(s)
#27-Dec-15 21:19

And to be clear, the branch model and the primitives model can both be expressed on the same set of vertices and object attributes - they can just be tables in between, linking which objects to which vertices in whatever way. There are different optimizations for different tasks of course, in OpenGL (and other 3d graphics I presume) the vertices are stored in a structural array, where "structural" means the indexing for primitives is a number-based offset into the array, rather than a DB-like ID "label". (I think this is the crux of the divide between GIS and gaming visualizations.)

I'll try to write this down somewhere in more detail, rather than pouring out here. :)


https://github.com/mdsumner

dale

510 post(s)
#28-Dec-15 02:45

And me, point clouds from SfM (Pix4D) are starting to flow my way.

dmbrubac


1,602 post(s)
#12-Feb-16 16:50

Here is a link to the proceedings of SilviLaser 2015. There is gold on almost every page for us lidar folks, but page 247 describes an algorithm I am particularly interested in: Generating pit-free DSMs.


Don't expect, suggest!

mdsumner


4,178 post(s)
#12-Feb-16 20:22

Thanks for this, I see that this was recently discussed on the LAStools blog:

http://rapidlasso.com/2016/02/03/generating-spike-free-digital-surface-models-from-lidar/

I'm going to try implementing this.


https://github.com/mdsumner

dmbrubac


1,602 post(s)
#14-Feb-16 18:30

Yeah that's how I found it too.

The paper is still under peer review, so we'll have to wait a bit to get the full methodology.


Don't expect, suggest!

Mike Pelletier


1,367 post(s)
#29-Dec-15 23:33

What a treat to come back to work and find this news. I'll just add that some classification tools would be a great addition to take advantage of the new speed. My little world involves the need for good vegetation mapping, both for wildlife habitat modeling and for creating attractive and informational (realistic) shaded relief. Currently available vegetation mapping is typically 30 meter which is only good for academics but is way too coarse for work done by land managers. 4-channel imagery is widely available at 1 meter pixels and we happened to have 1 foot imagery. Just need a good tool to crank through the massive amount of data.

Attachments:
Capture2.PNG

Dimitri

4,024 post(s)
online
#30-Dec-15 10:38

Beautiful PNG!

I'll just add that some classification tools would be a great addition to take advantage of the new speed.

As always... http://www.manifold.net/info/suggestions.shtml ...especially given adamw's comment about "500 million points isn't really scary" Whoa! That reminds me of the famous line from the movie Jaws... "We're going to need a bigger hard disk."

Just need a good tool to crank through the massive amount of data.

Radian should be just the thing for that. It will be interesting to hear how it works out for you.

Forest
502 post(s)
#30-Dec-15 21:44

What a great announcement! I am doing a lot of environmental research into long term ecological changes in the environment due to climate change and the new openness matters to me as all long term datasets will be in open formats, so the ability to work with these formats is really important. I am looking forward to some way of working with point clouds. Beyond that I hope the suggestion list gets a look over as I am sure that there are a lot of gems in there, a vast number of low hanging fruit such as importing gpx files. Open source was threatening to accomplish everything that Manifold promised and there is nothing wrong with that. My use of Manifold was declining to only the really hard tasks that only Manifold could do well. One thing that still value greatly is being able to double click the Manifold icon and being able to start work two seconds later. Convenience is a feature. Speaking of icons, maybe we can also get some generation icons inside Manifold as well?

tjhb

7,283 post(s)
#30-Dec-15 22:29

Nice post, full of useful detail, though what does this last comment mean?

Speaking of icons, maybe we can also get some generation icons inside Manifold as well?

dmbrubac


1,602 post(s)
#30-Dec-15 22:58

But Dimitri, hard disks are basically free, except for this one!

Seriously though, half a billion points doesn't make for a terribly large LAZ file. It's quite efficient and I expect you smart folks will be just as efficient.

I'm really happy and surprised at the amount of demand related to processing point clouds. It's a huge pain point in our industry and if you solve it or at least vastly improve what we have today, it will be bigger than when we learned we could buy a capable GIS for a few hundred dollars.

Mike, if you need better imagery I could come for a visit. Most of my work is 2.5 to 3.5 cm


Don't expect, suggest!

artlembo


2,784 post(s)
#31-Dec-15 02:32

If the tools, flexibility, and logic are there in the radian SQL, it might be worthwhile to have a code sprint for North American beta testers. Working together to implement some lidar processing libraries in SQL.

dmbrubac


1,602 post(s)
#02-Jan-16 21:27

I'd be happy to participate.

Assuming there is a beta forum, would we use that or something else? Even though it would be a bit of a subversion of proper use of this forum, I think the advantages to all would outweigh the disadvantages.


Don't expect, suggest!

tjhb

7,283 post(s)
#02-Jan-16 21:51

No subversion as long as it is not geographically restricted. (Art's suggestion applying to North America only is probably offhand and in any case unnecessary.)

artlembo


2,784 post(s)
#02-Jan-16 22:47

there is a beta forum that Tim and I are currently a part of (no secret there, as Mike Fisher acknowledged the two of us). So, I suspect all discussion will move to that forum.

As for my mention of North America, since I live there, I would participate in a local code sprint. It is certainly possible to do a code sprint online and worldwide, but the FOSS4g code sprints usually are a bunch of people working in a hotel conference room together. That would be fun to meet people.

I would welcome one in Australia, Asia, or Europe, but would likely not be able to fly there to participate.

A little premature at this point, but lets see how this goes :-)

dmbrubac


1,602 post(s)
#03-Jan-16 03:04

Ah ok. I didn't think you meant face-to-face, but I would consider it.

I propose we carve out a few threads for ourselves in the Beta to:

  1. See how many beta participants are interested.
  2. Decide what we want to accomplish over a number of sprints (I'd like to see the process continue past the beta if possible)
  3. See what we can do with the first sprint being solely on the beta forum
  4. Decide at that point whether it is best to continue online, schedule a face-to-face sprint somewhere, plan for a sprint post-beta so that others can participate too, or some other viable path.

Treat this as a strawman to be remorselessly torn apart.

P.S. when I wrote 'subversion' earlier I did NOT mean Subversion the revision control system. I tripped over that myself reading Tims reply.


Don't expect, suggest!

mdsumner


4,178 post(s)
#03-Jan-16 08:20

I like this idea, we really need some kind of multi-point 3d revolution. I'm keen.

I'd want us to discuss PDAL, plas.io, WebGL, D3, PostGIS and lastools at least to put relevant context on what we'd like to see. Is it worth starting a laundry list of tools / topics to hash out?

(Speaking of subversion, which is surely now hopelessly out of date, I use Github and github.io - it's so easy to publish now, integrated with your Git source control. )


https://github.com/mdsumner

adamw

7,031 post(s)
#03-Jan-16 19:21

Is it worth starting a laundry list of tools / topics to hash out?

Yes. :-)

mdsumner


4,178 post(s)
#10-Jan-16 11:22

This is not much, but worth sharing I think: some 3d ravings with R. The message is probably obtuse for many Manifoldians, what I don't have clearly in there is that GIS and graphics/gaming don't easily mesh - but they should.

I'm fleshing this out with the gris package, and the prospects of running Manifold from R via dplyr are rather exciting.


https://github.com/mdsumner

adamw

7,031 post(s)
#11-Jan-16 10:48

A question regarding dplyr (I am not familiar with it) - in this fragment:

src_postgres <- function(dbname = NULL, host = NULL, port = NULL, user = NULL,

                         password = NULL, ...) {

  con <- dbConnect(PostgreSQL(), host = host %||% "", dbname = dbname %||% "",

    user = user, password = password %||% "", port = port %||% "", ...)

  src_sql("postgres", con)

}

...is the code behind dbConnect(...) specific to PostgreSQL or does it use, say, ODBC, or maybe some custom tech with miniport-like drivers for various databases (and there is a driver for PostgreSQL)?

Added: I get that the point of the code in the article is exactly to expose a custom data source via some standard API, I am just wondering about what we could potentially reuse should we try to do that.

mdsumner


4,178 post(s)
#11-Jan-16 11:46

Hi Adam, it is specific to PostgreSQL (not via ODBC), but I don't think I can answer the second part fully - it's all drivers, but some of them are thin-wrappers around the real DB library. I've been wondering if there'll be a more direct way to leverage Manifold SQL that is not via ODBC - and if so it's very easy to rewrite the R code to use that instead.

Here's something of a guide under the R hood:

The current backend for PostgreSQL in R is via this package:

RPostgreSQL

dbConnect in src_postgres, is running this R-level code, with a .Call() to the C-library underneath:

'R

"conId <- .Call("RS_PostgreSQL_newConnection"

https://code.google.com/p/rpostgresql/source/browse/trunk/RPostgreSQL/R/PostgreSQLSupport.R

In the C source, "RS_PostgreSQL_newConnection" is here in line 201:

https://code.google.com/p/rpostgresql/source/browse/trunk/RPostgreSQL/src/RS-PostgreSQL.c

I'm working on src_manifold here, which is using the RODBC package underneath.

https://github.com/mdsumner/dplyrodbc

I have an example there with Access (just because nearly anyone can reproduce that and it proves it works) and then another one with Manifold, my real target. It was gratifyingly simple (for an experienced R programmer).

My src_manifold is doing the real work with RODBC:

https://cran.r-project.org/web/packages/RODBC/index.html

Ultimately, this is all R code that leverages the underlying DB connection - usually by C code built into the R package - it just simplifies it (at the user-level!). The ODBC stuff for R is not "DBI-compliant" (which just means it doesn't have a front-end for the DBI package in R). The DBI package aims to be the controller for all other DB packages in R. So dplyr talks to DBI, and DB packages (like RODBC and RPostgreSQL and RSQLite) extend the virtual methods in DBI. There's a fair bit of package-chaos and fragmentation because things have been evolving very quickly, but the dplyr-DBI scheme is looking very stable now.


https://github.com/mdsumner

adamw

7,031 post(s)
#11-Jan-16 12:00

Thanks for the links, they help.

Yes, it'll be possible to use Manifold without going through ODBC, although ODBC is fine as well, at least at first.

hugh
125 post(s)
#11-Jan-16 21:14

Thanks for info on RPostgreSQL. Hopefully Radian Studio will change things but until then postgis is where my big spatial data with many hundreds of table fields has to live. Any thoughts on how the Microsoft-R connection via http://www.revolutionanalytics.com might how we could use Radian products?

adamw

7,031 post(s)
#12-Jan-16 08:39

Radian Studio will be able to interoperate with PostgreSQL / PostGIS right out of the box. We will be able to interoperate with R as well, eg, via ODBC (see posts above). It will be possible to write a native module for R, too, and if we find we need to have such a module built-in, we'll do it.

Dimitri

4,024 post(s)
online
#12-Jan-16 16:31

Any thoughts on how the Microsoft-R connection via http://www.revolutionanalytics.commight how we could use Radian products?

Sure. For example, you can use them to add value to this proposition just like with so many other Microsoft activities. There are lots of great links and other fascinating info in that web site so when you search for the right key words ("GPU" for example) if you are into adding value you'll see what I mean. Radian is not a released product so I can't get into specifics in a public forum but if you want to play around with some of that you should apply to join the beta.

hugh
125 post(s)
#13-Jan-16 17:27

With Radian coming, this: http://www.zdnet.com/article/microsoft-delivers-free-version-of-its-r-analytics-server-for-developers/, and open-sourcing .NET, the Microsoft ecosystem seems a good place to have stayed with

dmbrubac


1,602 post(s)
#04-Jan-16 18:33

Subversion isn't that bad. I use VisualSVN in Visual Studio and the server on the back end. I'm the only user though.

I'm happy to talk about all that and more, though I don't know what plas.io and D3 are.


Don't expect, suggest!

mdsumner


4,178 post(s)
#04-Jan-16 21:09

Check them out:

Load a LAS file in the browser

D3, data driven documents

And for D3, especially this

I've got a document that goes into detail about driving WebGL from R, which explains some of the things I'm interested in but I've had to park it for now.


https://github.com/mdsumner

dmbrubac


1,602 post(s)
#05-Jan-16 15:27

The only problem I see with that is the time needed to study, understand and then make use of it. My learning curve needs some adaptive resampling.

Thanks Mike!


Don't expect, suggest!

dchall8
289 post(s)
#28-Dec-15 22:45

I'm sure we're all anxious to see what's new. I don't really do anything to challenge the capability of Manifold 8, but I am sitting on 800 square miles of LiDAR data that I'd love to process. We're looking for buildings hiding under the trees in the county.

Will Manifold 9 have the same GUI as 8? One of the problems ESRI had in the past was they changed their menus and toolbars from generation to generation. That required their users to go back to ESRI college simply to continue to work. Can we expect Dr Lembo to put together a Manifold 9 video series??

Our budget for the year is in. Hopefully we'll see pricing by mid year, so I can get this into the office budget for 2017. That is, of course, assuming this won't be one of the old $50 upgrades from the past.

adamw

7,031 post(s)
#29-Dec-15 10:02

Regarding the UI - it is not the same as the UI of Manifold 8, but it is similar enough so that a long-time user of 8 likely won't have too many difficulties adopting. Think Windows 7 vs Windows XP (not Windows 8 with the tiles and the start screen instead of the start menu, etc, vs Windows 7).

dale

510 post(s)
#27-Dec-15 00:43

From the announcement "This design feature enables easy creation of new applications and new features both by Manifold and by users. Manifold expects to utilize Radian configurability to issue many updates to Release 9 in 2016.

Looks like it's time to revisit suggestions, now that functionality improvements will be back.

As an aside, what great and welcome news.

Gustavo Palminha

1,009 post(s)
#27-Dec-15 01:56

Good news for the upcoming year and looking forward for the beta testing.


Stay connected:

Linkedin >> http://www.linkedin.com/in/gustavopalminha

Twitter >> http://twitter.com/gustavopalminha

Other >> http://www.spatialmentor.com

mdsumner


4,178 post(s)
#27-Dec-15 07:14

It's been a lot of christmasses. A glorious 8 year reign. :)


https://github.com/mdsumner

adamw

7,031 post(s)
#27-Dec-15 08:02

Looks like it's time to revisit suggestions, now that functionality improvements will be back.

Absolutely.

danb


1,573 post(s)
#20-Jan-16 19:36

First up great to hear that Manifold is still evolving and moving towards the release of Radian Studio and Manifold 9.

For those of us wishing to be fully prepared to make the most of Radian technology, I was wondering if Manifold are planning to indicate a few broad guidelines for a 'Radian ready' machine tailored towards making the most of the new technology? Clearly this would be out of date immediately if it was to list actual specs, though in the past, I have found some of the more general tips such as dedicated SSD for OS etc provided by Tim and others to be enormously valuable. This sort of information is currently scattered across a number of threads, each littered with out of date specs. I'm not sure if a it is planned to add a dedicated Radian section to the forum, but this could perhaps sit in one of the FAQ sections of the forum as an ongoing reference?

Such guidelines might be aimed at different levels of usage and for differing intended dominant usage ie lidar and grid processing, Spatial SQL etc (where a dominant usage warrants a hardware bias with Radian). I'm sure that much of this might be common knowledge, but I find myself revisiting this periodically and I think that it might be useful to have a resource containing some definitive guidelines and an indication of potential pitfalls in commonly tendered prebuilt specs.


Landsystems Ltd ... Know your land | www.landsystems.co.nz

Dimitri

4,024 post(s)
online
#20-Jan-16 20:05

To do a good job advising on hardware you really need to be able to discuss in detail the planned use, the trade offs with constraints such as budget and facilities like space available and then you also have to be able to talk about the details of the full stack of software you plan to use. Can't do the latter in connection with Radian outside of the beta program until the product is released, but once it is I'm sure many people will have a lot of views on what works best for them.

My own two cents is that hardware ran into the end of Moore's Law long ago so it hasn't changed all that much and strategies therefore tend to be the same: run 64-bits, put your money into lots of memory before making a diminishing-returns expense on a slightly faster CPU, make sure you have an NVIDIA GPU (just my two cents...) without going nuts on diminishing-returns expenses for the absolutely most psychotically latest GPU and so on.

I find that what counts a lot more for me in daily use productivity is ensuring I have lots of desktop real estate (I always configure machines I use with three monitors), I want super reliable machines (with disks always mirrored for reliability) , I get solid and easy backup technology and a great UPS. To me that's more valuable for productivity than spending the same amount extra on a slightly faster CPU but trying to live with a single monitor, un-mirrored storage and a feeble UPS.

danb


1,573 post(s)
#20-Jan-16 20:23

Many thanks Dimitri, fair comment regards commentary outside the beta prior to release and useful other comments which I will no doubt refer to in the future.


Landsystems Ltd ... Know your land | www.landsystems.co.nz

jburn

540 post(s)
#05-Feb-16 23:58

I'm certainly pretty excited about this prospect. Really looking forward to it!


__________ jaburn

Dimitri

4,024 post(s)
online
#06-Feb-16 07:09

make sure you have an NVIDIA GPU (just my two cents...)

To amend that... one guideline coming out of various betas is that a Fermi or later Nvidia GPU is required. This is talk about fairly old hardware but it still may be applicable.

Fermi is compute level 2.0, a big step up by Nvidia from the prior CUDA technologies at 1.x or lower. Radian requires at least compute level 2.0, so Fermi or later GPUs are required.

This should not be a practical limitation for most. I'm writing this from a home system which has a couple of super-cheap and very old GPU cards, so cheap and so old I don't even remember what they are (ah... wait! There's an NVIDIA Control Panel for that... they're GeForce GT 520's), but the only reason they are in there is to provide two cheap monitor connections per card with DVI plugs so I can run three monitors.

I think such cards sell for... what? $50? less? ...these days and yet they are still GPGPU performance enhancers with Radian at Fermi or later levels.

Older systems with lots of RAM if they were hot systems to begin with still have razzle-dazzle (an older Core i7 with tons of RAM and a couple of terabytes of RAID disk is still plenty fast). The only issue you tend to run into with older systems running more recent plug-ins like newer GPU cards, etc., is the older style BIOS the older systems tend to have, with which the new cards won't work.

So to plug in a newer GPU if you can't find a version that will also work with older BIOS you end up having to buy a new motherboard with a new socket for a new processor and new RAM that aren't noticeably better than the old motherboard with the old socket with the old processor and tons of old RAM you had. As much as I like having the latest gadgets I feel badly about swapping out a perfectly good motherboard/CPU/RAM ensemble in a home system that's at best lightly used in any event. Such is life in the fast lane of technology...

mdsumner


4,178 post(s)
#06-Feb-16 11:49

Thanks Dimitri this is very pertinent. Perhaps worth pointing out that 8.0 does report "Fermi" or "pre-Fermi" for cards in the Help/About dialog.


https://github.com/mdsumner

KlausDE

5,994 post(s)
#27-Dec-15 08:26

I haven't seen a restriction to a special type of GPGPU for testing. Correct?

adamw

7,031 post(s)
#27-Dec-15 08:59

No restrictions, but analysis specifically will only benefit from NVIDIA for now.

KlausDE

5,994 post(s)
#27-Dec-15 10:30

Thx, so we are prepared. There has been a warning for long that only newer versions of windows will be supported in future. Is Win 7 OK?

adamw

7,031 post(s)
#27-Dec-15 10:38

Yes, Windows 7 is supported.

(We'll probably support it for a long, long time, it was a great version of Windows.)

KlausDE

5,994 post(s)
#27-Dec-15 10:50

Given the short time planed for the beta I hope that a help system is in preparation. Will we be able to benefit from a documentation of the object model at least?

adamw

7,031 post(s)
#27-Dec-15 10:58

There will be technical documentation (ie, for the object model, a brief description of all objects and methods, but little in the way of overviews). The full help system is in preparation, yes.

The ideology is different from Manifold 8 or prior (way more open) though, and that helps a lot.

cartomatic

884 post(s)
#05-Jan-16 19:37

Looks like lots of goodies is on the way. Can't wait too :)

How about the IMS? Is it exposed in the API already? OGC compliant WMS / WFS?

Will radian / manifold be able to render transparent antialiased images?


maps made easy - www.cartomatic.pl || www.cartoninjas.net

adamw

7,031 post(s)
#06-Jan-16 07:47

Some parts of what you are talking about will happen in Radian Studio, other parts in Release 9.

Come to the beta for details.

(Just to be sure, by the last sentence you mean rendering screens with antialiased vector data and transparency programmatically for further composition - eg, sending over the web via OGC WMS - correct? If it's something else, please clarify.)

cartomatic

884 post(s)
#06-Jan-16 08:34

I have applied for the beta so hopefully will get there too.

Regarding the transparency - yes i am after being able to render some vector or raster on a map with transparent background so such an image can later be overlayed on other layers without having to remove background in own code.


maps made easy - www.cartomatic.pl || www.cartoninjas.net

DomagojC6 post(s)
#26-Feb-16 11:13

I don't know if you are registred me as Release 9 beta tester?

What i must do to be a Release 9 beta tester?

cartomatic

884 post(s)
#11-Jan-16 09:25

How about improving data linking - specifically linking ecw in code. Has this been addressed by any chance?

Also - this has just come to my mind - ability to specify whether the linked components refresh on open or not, when opening a project file in code. This would be very useful in IMS too.


maps made easy - www.cartomatic.pl || www.cartoninjas.net

adamw

7,031 post(s)
#11-Jan-16 10:43

We do have improvements in that area (linking data). I can't go into detail in this thread though, my apologies.

cartomatic

884 post(s)
#11-Jan-16 11:23

glad to hear that :)


maps made easy - www.cartomatic.pl || www.cartoninjas.net

Dimitri

4,024 post(s)
online
#11-Jan-16 12:30

Also - this has just come to my mind - ability to specify whether the linked components refresh on open or not, when opening a project file in code. This would be very useful in IMS too.

As always, http://www.manifold.net/info/suggestions.shtml ...the more, the merrier!

cartomatic

884 post(s)
#11-Jan-16 13:12

Obviously. Thanks for the reminder. Sent.


maps made easy - www.cartomatic.pl || www.cartoninjas.net

scandalxk53 post(s)
#14-Jan-16 14:18

Responses to this announcement are biased to the techy end of the user spectrum. I, on the other hand, am an environmental consultant who needs a reliable, economical, easy to use GIS and never customises it or attaches it to external databases or publishes interactive online maps or loads terabyte-sized databases or anything particularly challenging at all, really...and I am very, very pleased to hear that Manifold is moving forward.

Good luck with the rollout.

dchall8
289 post(s)
#15-Jan-16 16:47

I predict someday your environmental consultancy will be interested in LiDAR analysis. LiDAR is an imaging technique where you can distinguish among reflections from the ground, buildings, and plants. Accuracy is measured in fractions of a foot heading to fractions of an inch.

An interesting application is to measure California's water resources. The snowpack in the California mountains is measured by subtracting the winter LiDAR scan from the summer LiDAR scan. The difference is the exact volume of snow.

ColinD


1,784 post(s)
#15-Jan-16 19:44

Similarly, as an ecologist I use LIDAR to detect surface changes following longwall mining panel subsidence, particularly hydrology changes.


Aussie Nature Shots

scandalxk53 post(s)
#18-Jan-16 15:57

Thank you dchall8 and ColinD. Yes, I have used LIDAR, mainly for generating ZTVs in forested or built-up areas where the DTM does not equal the visible surface.

adamw

7,031 post(s)
#16-Jan-16 08:04

(Bumping this thread a bit after the influx of bots.)

firsttube


1,427 post(s)
#18-Jan-16 13:49

There were quite a few interesting suggestions and feature requests at the user conference in May 2008, Salisbury, MD if I remember correctly. I know we can't comment specifically on those, but I'm hoping some of them were, or are going to be, incorporated into the next release (?)


"The blessing in life is finding the torture you are comfortable with." - Jerry Seinfeld, 6/26/2013

adamw

7,031 post(s)
#18-Jan-16 16:09

We definitely did a number of things based off those requests, yes.

The best way to check for specifics is to come to the beta (sorry for repeating the obvious).

DomagojC6 post(s)
#17-Feb-16 12:43

Hi!

Can i be4 a beta tester too?

mdsumner


4,178 post(s)
#17-Feb-16 12:49

Please see this thread for details on that (it's linked on the front page of the forum currently):

http://www.georeference.org/forum/t129214

FWIW, none of the participants here can tell you anything about the beta, whether on it or not, other than pointing to that thread. All the best.


https://github.com/mdsumner

sgaiji20 post(s)
#11-Aug-16 03:05

Hi all,

How we will be advised that the new version is available for buying and downloading ?

Directly from Manifold when it try to detect if there is a new version ?

Thanks.

Stephane

Dimitri

4,024 post(s)
online
#11-Aug-16 05:18

See http://www.georeference.org/forum/t129214

cab68 post(s)
#25-Aug-16 07:04

...after Radian Studio, most likely in the first half of 2016.

When do we expect to purchase Radian Studio?

We need to spend our budget next month...

Dimitri

4,024 post(s)
online
#25-Aug-16 12:38

Note from http://www.georeference.org/forum/t129214 "Manifold cannot provide any additional public information on either Radian Studio or Manifold System Release 9 prior to release of these products. When these products are released the web site will provide detailed information. "

It's a compliment, of course, if you want to give Manifold your money without knowing what the product is, but my advice would be to learn about it first. The way to do that if you are in a hurry is to participate in things like the beta program. Did you apply to join the beta?

cab68 post(s)
#26-Aug-16 03:38

Thanks Dimitri! I applied as a beta tester, unfortunately no reply from Manifold whether I qualified or not.

Dimitri

4,024 post(s)
online
#26-Aug-16 09:25

unfortunately no reply from Manifold whether I qualified or not.

That's very puzzling as I had heard absolutely all inquiries got a reply as described in the announcement.

Perhaps I misunderstand what you mean, but if you mean a reply "you're qualified / not qualified" that is not what should have been expected. To clarify:

Visit http://www.georeference.org/forum/t129214 ...note this part:

Tech hopes to acknowledge all letters but that may take a few days or a week or more. If you are selected for participation you will be notified by email. If you are not selected you will not receive an invitation but will not receive a decline.

I heard that tech did acknowledge all letters. Did you get such an acknowledgement? If you did, the content of that acknowledgement should have been clear.

If not, did you follow up with any messages to Manifold?

From what I've heard, every case where someone followed up to write "Hey, I applied and never got a response" was tracked down to one of two issues: a) they got a response that their receiving email system shunted off to a junk mail folder or b) they didn't really apply.

For example, the instructions say to...

please write from the email address used for your Manifold license to tech support at the published tech email address with a subject line that includes "Radian Studio beta test" with a statement that you would like to participate.

... if someone ignored the above, well, try as tech might to detect all inquiries it's certainly possible whatever message was sent was not detected as a beta program application, and thus no reply from the beta program was sent.

It's a big deal if tech didn't respond as set forth in the announcement so this is worth a bit of clarification. If tech made a mistake that needs to be corrected.

cab68 post(s)
#29-Aug-16 11:23

Hi,

Actually, there was reply asking for my personal email address, but after giving my email address there was no reply anymore.

Anyway, we will just wait...

Dimitri

4,024 post(s)
online
#29-Aug-16 15:27

Actually, there was reply asking for my personal email address, but after giving my email address there was no reply anymore.

It looks like your email server did not deliver tech's reply to you.

When you mentioned a few days ago you did not receive a reply I asked around. Tech did a full audit of all beta communications to find all situations where it was clear people who wrote about the beta had problems receiving Manifold's replies. What you describe above is one of those cases.

The problem is that the email address used goes through a server that sometimes delivers Manifold replies and sometimes not. Sometimes it sends an error message that it could not deliver a Manifold reply and sometimes it does not. Other people in your organization also are apparently using the same email address. When someone else recently wrote tech sent five replies and had all five rejected by the server.

The solution is simple: do not use your organization's generic email address or email server to communicate with Manifold. Instead, get your own email address, like a gmail address, and always write from your private email address to communicate to Manifold.

See the page at http://www.manifold.net/tech/email_problems.shtml

cab68 post(s)
#29-Aug-16 23:55

Thanks for the info. I did not use my gmail or yahoo email address.

Dimitri

4,024 post(s)
online
#30-Aug-16 08:07

No worries. I'm glad we got this solved.

It is up to you what you do from here, but my advice would be to follow up with tech from your gmail or yahoo address. If tech reached out to you with a second effort regarding the beta, they want to get you or your organization involved, if not now than maybe later.

If the beta is no longer of interest to you it still would be a good idea to establish reliable contact with tech, in case that might be useful to you or to your organization in the future. :-)

sgaiji20 post(s)
#04-Jan-17 10:00

Happy New Year all !

Is there any fresh news about Radian or/and Manifold next release ?

Thanks

Stephane

sgaiji20 post(s)
#25-Jan-17 07:37

It's been more than 10 years since I fell in love with Manifold even though I still have a lot of difficulties writing scripts or SQL queries. It is now 5 years that I look forward to the new version.

For 1 year, all my clients ask me to deliver my work in QGis format (I continue to deliver my works in both Manifold & QGis format).

I asked at the beginning of the year news about the development and the availability of this new version without answer until today.

I repeat my request in the hope that a charitable person can reassure me and convince me that this will take place soon ...

A basic user

Dimitri

4,024 post(s)
online
#25-Jan-17 10:51

ask me to deliver my work in QGis format

Shapefiles? Are you serious?

Stay tuned soon for a Radian release near you. Days, not weeks.

mao9021048 post(s)
#26-Jan-17 06:45

looking forward about that.. can i a have a trial version? and if not how much will cost me?. thank you

i am using manifold for 7 years and still i am a basic user..

cab68 post(s)
#26-Jan-17 07:21

I'm so excited!

mechalas

838 post(s)
#27-Jan-17 20:21

Shapefiles? Are you serious?

What file format would you suggest for exchanging geometry layers with other applications?

Dimitri

4,024 post(s)
online
#28-Jan-17 08:23

Not the right thread to pursue a discussion of shapefiles.

mechalas

838 post(s)
#30-Jan-17 03:01

If 9 expands the drawing export options then I will be happy. With 8, KML/KMZ and Shapefiles are the two formats I can export to that just about everyone can deal with in some fashion, and of course neither is ideal.

tjhb

7,283 post(s)
#30-Jan-17 05:04

John,

I don't know how you managed to be excluded from the beta. Just timing perhaps.

You'll be pleased and impressed with what's been done.

It will require new learning of course. A very different thing than 8.

mechalas

838 post(s)
#30-Jan-17 05:42

I have not been involved in any GIS projects in the last few years, so haven't been very active here or elsewhere. Only recently have I had the need, and it's been...distinctly weird.

tjhb

7,283 post(s)
#30-Jan-17 05:55

That explains it.

There is a tremendous amount to learn. Most of it is new concepts.

mechalas

838 post(s)
#30-Jan-17 12:43

I am pretty excited about this, and looking forward to it! And Dimitri's comment about days rather than weeks is more than encouraging.

mechalas

838 post(s)
#30-Jan-17 15:29

Just ordered my copy of Radian.

antoniocarlos

444 post(s)
#30-Jan-17 16:05

Introductory offer question.

Buy Radian Studio Universalbefore15 February 2017and get a Manifold Release 8.00 Universallicense at no charge. For each Radian Studio license in your shopping cart the Online Store will automatically add a Manifold Release 8.00 Universal license at no charge. This offer ends at midnight GMT on 15 February 2017.

If I purchase Radian, can I install it on my current version of manifold 8 and install the extra license on a second computer without Radian?

ACGT


How soon?

mechalas

838 post(s)
#30-Jan-17 16:13

Yes. You get a separate serial number for Manifold 8.

adamw

7,031 post(s)
#30-Jan-17 16:17

Be sure to buy the bundle product (Radian Studio 9.00 Universal x64 + Manifold 8.00 Universal x64), not just Radian.

It should be the topmost item on the store, etc, and maybe I am stating the obvious, but better safe than sorry.

ColinD


1,784 post(s)
#30-Jan-17 20:45

There is a discrepancy in the closing date for the combined offer on the web site. The front page says closing 12 Feb 2017 while the store says 15 Feb 2017.


Aussie Nature Shots

adamw

7,031 post(s)
#31-Jan-17 06:50

Thanks for the note.

I asked and the correct date is 15 Feb. The web site team say the wrong "12 Feb" has been fixed.

ColinD


1,784 post(s)
#31-Jan-17 07:11

Not quite yet unless it just hasn't propagated

Attachments:
31-Jan-17 6-08-44 PM.jpg


Aussie Nature Shots

adamw

7,031 post(s)
#31-Jan-17 07:18

Heh. Sending it to the team immediately. Thanks a lot.

jockeryl
170 post(s)
#31-Jan-17 08:25

Is there any information about upgrades from Release 8 to Release 9? Can we expect that there will be an upgrade license path as it was between 7 and 8?

I have sold around 20 licenses of R8 and would like to be prepared to upgrade and train the users and if possible be able to budget the upgrade for my customers in advance while evaluating the new features.

I understand that Release 9 is not yet announced and it will be 'later this year', so any indication in regards to upgrade offers at this point will be useful.

For now I will start learning Radian.

Dimitri

4,024 post(s)
online
#31-Jan-17 12:37

Is there any information about upgrades from Release 8 to Release 9?

No, but I can't imagine Manifold would issue 9 without some sort of special initial offer for 8 licensees.

orerockon

336 post(s)
#31-Jan-17 23:12

So can I assume that will be a limited time offer? Can anyone think of a way that one could get wind of it without checking this site or the website often? Maybe somewhere we could sign up for an email notification?

Dimitri

4,024 post(s)
online
#06-Feb-17 14:05

So can I assume that will be a limited time offer?

? My apologies for being unclear. The answer to your question "is there any information about upgrades..." is there is no information about upgrades from Release 8 to Release 9. If there was it would be on the web site or here as an announcement.

Beyond that, I think anybody who's noted how Manifold has rolled out new products in the past would make the same comment: they'd be surprised if some sort of initial offer wasn't made.

On staying in touch: I don't use RSS but I suppose you could use that (note the icon at the top of this page) to avoid manual checks. In theory, the automatic updates notification of 8 could be used for a notification of 9 as an update. I think it was used in the past when a new edition came out for the GIS. But that's no guarantee for the future.

Emails: The company has historically been extremely reluctant to send any sort of unsolicited email. There's an experiment going on right now where some emails are being sent out that are very brief and not intrusive. But that's as far as they'll go.

orerockon

336 post(s)
#08-Feb-17 22:00

Open open open open... :D (For those of you who remember silly US TV commercials from the previous century.)

chouse
65 post(s)
#03-Feb-17 18:46

This was curious to me. I was wondering is there a benefit to me to upgrade to Radian now taking advantage of the universal license upgrade OR wait until M9 comes out and upgrading entire system. I read somewhere in this thread or another different thread a "comment' about radian would most likely be the work horse behind all future Manifold products. If I read into that comment and make an assumption that Radian could/would very well at the very least be bundled with future M9 product later this year.

I do just as much production maps and cartographic work as I do database work. I use postgres/postgis mostly so I do spatial in the database as much as in a GUI.

I currently have Manifold 8 Enterprise Edition with no extensions (i do most of the work of extensions by other means at the moment...though I like diversity of options to explore).

Thanks for any advice. This is kind of a What would you do question than anything.

tonyw
379 post(s)
#03-Feb-17 19:34

The announcement on Radian Studio and Manifold 9 suggests "Experience with Radian Studio will directly translate into a faster start with Release 9" which is why I purchased RS with the hope of being able to put Manifold 9 to work faster. I'm not a power user at all, I don't program anymore or write script and I don't have a database. If you already work with databases maybe the ability to get into M9 faster is worth getting RS9?

Dimitri

4,024 post(s)
online
#06-Feb-17 14:11

radian would most likely be the work horse

From the Announcement in this forum

Given that Manifold Release 9 is based upon Radian technology, using exactly the same engine, formats and localization files as Radian, we expect Radian users will be able to seamlessly upgrade to 9.

Great care has been taken throughout Radian development and will continue to be applied to ensure that from a user's perspective Release 9 will be a seamless upgrade and superset to Radian. For example, the native format Radian uses is known as Manifold 9.00 .map format.

If interested in 9, get Radian and build expertise in Radian. Early access for 9, for example through the beta test program, will be offered only to Radian licensees.

A lot of this boils down to budget. If you have the funds and you are working with databases I can't imagine you wouldn't find Radian useful.

tomasfa
130 post(s)
#30-Jan-17 21:18

Getting my Radian, in 3, 2, 1...

tonyw
379 post(s)
#31-Jan-17 06:13

Those who ordered Radian, how long before the serial numbers arrived in a follow-up email? I ordered about 13 hours ago, immediately received the confirmation of the order but have not received the serial numbers yet.

The confirmation email read: "Your order with manifold.net has been processed. Your serial number email with download instructions has been sent to ______. Print the invoice below for your records."

Thanks

dale

510 post(s)
#31-Jan-17 10:33

Website is explicit, you should expect immediate delivery of serials. The email hasn't ended up in your spam folder, or blocked by a corporate filter?

cab68 post(s)
#31-Jan-17 11:30

At last! it came out. What a longgggg wait!

Let's hurry up for the Promo $395 Radian +Manifold 8 (64bit)

orerockon

336 post(s)
#31-Jan-17 23:09

Egads I just now decided to check if there was a new release after not checking for a couple months. Stranger things have happened! As a amateur user I'll be waiting for release 9 though.

tonyw
379 post(s)
#31-Jan-17 17:18

Nope, not in spam/junk folder. I have several PCs and my cellphone checking the same account and a backup gmail account that checks the account where the serial numbers should arrive. No sign of the serial numbers. I'll check with sales.

Dimitri

4,024 post(s)
online
#01-Feb-17 07:26

checking the same account and a backup gmail account that checks the account where the serial numbers should arrive

Ohh... cannot resist. There's a web site for that! :-) Email Problems

The boldfaced Important bits on that page are especially on point in this case: if something in your account or your ISP is blocking inbound email it doesn't matter how many different gadgets you use to check that account because the account itself is the problem, not the browsers on your gadgets.

It also doesn't help to use gmail to poll a defective account or to use a defective account to poll gmail because.... well, the account itself is the problem. You have to use gmail directly and in the default configuration to avoid account misconfiguration from blocking you from receiving legitimate emails you want to get.

tonyw
379 post(s)
#01-Feb-17 19:03

Thanks Dale,

I worked with Manifold sales and incrementally backed off the aggressivity of spam filters and eventually disabled the spam filters but that didn't work. I own the domain name and control the email accounts through settings in the web hosting subscription. I then gave Sales a gmail account and the serial numbers came through on the first try. For some reason while I could communicate and receive emails from Manifold Sales using the same company email address, the serial numbers would just not come through, not even ending up in spam/junk. Whatever filtering is happening is beyond what I have control over. If anyone else is having a similar problem receiving the serial numbers, providing Sales with a gmail address works.

dale

510 post(s)
#02-Feb-17 06:32

Excellent, and I'll be sure to use a gmail address rather than my usual (over which I have no control)

Manifold User Community Use Agreement Copyright (C) 2007-2017 Manifold Software Limited. All rights reserved.