Posts made by toxicvoxel

http://www.flickr.com/photos/luis_ruiz/

There are a number of anoying trends that go along with this kind of economic climate.(I have been arround long enough to see a few cycles.)

1.] The first people that get hit are the contractors as companies will try and hang on to their staff base in the downturn. Additionaly there may be legal issues that may arise if contractors are used while a company is anticipating or implementing redundancies.

• The good news is that when work starts to pick up again many companies will tend to use contractors in the first instance instead of taking on staff until such time as a consistent pattern of work becomes apparent. Unfortunately this trend may still be some time off in the UK and the US.

2.] When people go out of work the natural reflex is to spend a lot of time in building out additional skills in order to be more attractive in the market place. One of the typical target skills at the moment would be to develop CG rendering skills, so expect there to be a lot more competition in this market after the upturn. (The problem is that only a small percentage of clients can really differentiate this service based on quality.) Also expect a lot of purist cad monkeys to re-surface with Sketchup skills in the next 18 months so beware if you percieve yourself to be safe as being the 'Sketchup guy' in the office.

This is an inherent disadvantage of any IT based skills because what starts out as a specialist service will in time become a general skill or become redundant as new technology is adopted into the mainstream. (Eg. CAD > BIM, AutoCAD/MAX 3D > Sketchup) Now more than ever, one really needs to plan your investment in IT (self-)training / expertise carefuly in order to ensure your relevance and logevity in the market.

Skills that require the 'signature' or the hand of the author eg. traditional rendering (even if this is a small market) will be harder to replace with a cheap young graduate just out of school and will be less affected by changes in technology. Whatever you do needs to focus on the possibility of providing, where possible, a unique and customised service instead of competing with a growing supply of the standard service in the same field. If you don't all you will be able to do is try and compete on price, which is the last resort before going out of bussiness.
Part of the longer term solution is also to try and anticipate the IT workflow requirements in 5 years time and try and position yourself accordingly in terms of adopting emerging technology and developing skills that may be in shorter supply in future.

Your importer will have to handle faces on the world plane as SU always reflips these on import.
http://forums.sketchucation.com/viewtopic.php?f=180&t=31887

Some thoughts you may want to consider:

1.] Focus on developing a cost efficient design and develop a clear cost structure with the architect at the outset instead of focusing on saving on architect's fees.
If you find an architect willing to work for substantially lower fees chances are that he will not be generous with his input and what you save on fees may invariably translate into a less efficient design both in terms of initial cost and long term running costs.( I remember a client of my old business partner who came back after rejecting our fee to show me the plans that he had done for about 20% of our quote. When my partner did the cost calculation on the floor space taken by the long corridors in this design and showed him comparable circulation space in our designs he realised that what he had saved on our fees would eventually cost more in the cost of the building.)
2.] Make sure that you know who is actually designing the home. Is it the architect you interface with or a junior in the back-office doing the real 'design' work? From this point of view consider the size of your project relative to the general workload of the practice. A very prestigious practice may relegate your project which may be relatively small relative to their other commitments to junior staff. Find a practice where the size of your job is relatively substantial when compared to their other projects and preferably one that specialises in home design. Sometimes a practice may have a specialist designer who actually does the design work. If so ask that he be involved in discussions relating to the brief so that the message is not lost between the bloke doing the interface and the design department.
3.] Ask the architect to do one or two workshop meetings where you bring as much information and examples of what you like for discussion before any design work is done. This will help (and force) the designer to focus on your needs before embarking on a design which may not be satisfactory. These discussions often help you understand your own needs and priorities. Develop a clear written brief before instructing the architect to proceed taking special note of your requirements in terms of fittings and finishes as this aspect often causes cost overruns because they are not clearly defined. This way you will have a way against which to measure the performance of the architect (and your own expectations) in the event that there is a dispute or you are unhappy with the design. He may charge a few hours for this but it may be the best money you spend on the whole project.

If the architect feels that he cannot accommodate any of the above [2 & 3] find someone else.

You need to ask yourself some questions:
What happens to the drives (containing your data) on these servers when they are decommisioned , the service goes bankrupt or if the hardware is stolen?
Do you know the poeple who run this company? (Would you give a CD-Rom of your most sensitive data to someone you do not know to keep?)
Would your clients be happy in the knowledge that their data and designs (for example banks, prisons, private residences etc.) are being stored in this way?
If your backup contains any personal data you have obligations under the Data Protection Act (UK) in terms of how the data is encrypted and secured. You will not have a defense if these obligations are breached in the context of storage on-line.
I do not think this 'spreads' risk but rather increases it.

I have been looking at tablets lately from the perspective of using it as a portable painting device for use with Corel Painter.
Found this video relating to the new Asus Ep121 Slate which is quite interresting:
http://www.sluniverse.com/php/vb/science-tech/54734-asus-ep121-slate-vs-wacom.html

@unknownuser said:

..and I would be hard pressed to see the Sketchup beneath.

For me one of the major advantages of NPR over GCR is the fact that you do not always need a detailed model on which to base an image. You can develop early design study visuals based on your (or your client's) design sketches and other image references before the design has been finalised. In this way NPR can feed into a design process at a stage when CGR cannot.

If you are looking for a 'technique' type tutorial I'm going to have to dissapoint you because I dont have one. Rather the principles descibed above need to be understood and applied in whatever way that fits the image you wish to achieve.

I think the whole point about doing NPR is producing images that have a personal feel and show your own involvement as apposed to the machine perfection of GC rendering. Simply following a technique to achieve a specific style seems to be a little counter-productive to his goal. Instead of searching for 'recipies' I think that learning the nature of 'ingredients' that contribute to an image and how these can combined in different combinations would be a more helpful way to approach this.
This keeps the process flexible and spontaneous and will help you to develop a personal 'look and feel' to your images.

One tip I can give you is to paint in greyscale/monochrome in order to develop a feel for the tonal structure of an image.
Achieving both colour & tonal control at the same time is difficult when you start out. If you can't achieve a reasonable monochrome image you will struggle with images that contain mid tonal colour.

As with CG rendering there is no quick way to become more proficient so you do need to spend the time to develop your personal technique.

Not sure what subset of functionality P. Esentials has but I suspect you will have everything you need without spending on the full version.

[As far as being a master I'm not so sure as I still seem to end up with more abandoned images than completed ones.]

I think the key is not to start with Painter but rather the style you wish to replicate. Study the 'mechanics' of the mark-making and then select/customise brushes that give you the closest effect at the pixel level. If you want to replicate watercolours you need to understand the layers that have a transparent algorythm for achieving the look of overlaid washes. (Gel, overlay, screen, colorise.)

Guache / oil type styles are in my opinion a lot easier (albeit a little harder in 'look and feel') as you dont need to master transparency and can often do the paintwork on a single layer:

Use a texture overlay rather than expressing the paper texture as this will yield more realistic results.

Don't rely too much on step 'techniques' as these can generaly be easily identified from those where actual hand brushwork is involved. Use a tablet even if only a small cheap one.
Try and blur the line between traditional chemical media and digital.

PRODUCT UPDATE 2011/2/18
I have uploaded an update release (1.0.2.0) which provides additional import support for the free version of Sketchup via ruby geometry scripts and a friendlier installation experience for Windows 7 users. This is a free upgrade.

http://forums.sketchucation.com/viewtopic.php?f=80&t=31631

Essentialy done with Gel layers & Digital watercolour and Drip brushes.

Have been experimenting in replicating looser styles in Corel Painter.

http://gfxspeak.com/2010/11/12/autodesk-letting-out-teasers-about-new-architectural-concept-modeler/

The problem is that once you have come to a decision to move you may be knee-deep in a dispute with your hosting provider who can often be fairly un-cooperative in these circumstances. This can make it tricky to get your existing domain released/transferred and get up and running elsewhere.

A strategy that will enable you to change hosts in a couple of hours would be to purchase your domain url through an entity un-associated with your hosting space provider.
You simply point your domain to the hosting space url of your current provider and you never have to struggle with getting the domain transferred. When things go wrong you can move your website content to a new provider and just point your domain to the new provider. You can do this in an hour or less with no downtime because re-pointing can be done in minutes from a control panel without losing any links or email addresses once you have your website re-loaded on a secondary server.

This is also a good set-up if you have to reload a new site design from time to time. You can upload the new site to an alternate web folder and just point the domain to it when it is ready to go. If your website is critical such a set-up will allow you to have a second dormant clone of your website hosted elsewhere to which you can point your domain if the primary hosting server fails.

The domain pointing company that I use to do this is http://www.123-reg.co.uk but there are probably others that will allow you to use the same strategy.
The control panel provided by their website also allows you to point domain email addresses to any others your may use.

Looking into it further it turns out that add_face on the ground plane appears to re-organise vertexes using rules which would be complex and time consuming to test if vertex indexes are required to set edge visibility. A possible workaround would be to create a temporary group , add 10 to the z values of the face vertexes and then transform the group to the correct position where it can be exploded.

The problem with the Ruby API is that it requires so many hacks & workarounds. It would be in Google's interest to open the API development source so that it can be corrected where necessary and extended to address shortfalls. It is difficult to accept that errors in 7.x documentation remains uncorrected in 8.0.

I looked at .add_faces_from_mesh before (which does not follow the 'ZeroZ' rule by the way), but because it has poor support for controlling individual outer (naked) edges of the mesh it's use is limited. If you have to iterate through the entities collection anyway to find and set individual edges it completely negates the efficiency of the method.

Is it just me or is the documentation fraught with errors? The method does not return a collection of faces as stated:

Syntax
faces = entities.add_faces_from_mesh polygonmesh
Arguments
polygonmesh - a Polygon mesh object
Return Value
faces - an array of Face objects if successful 

Thanks for your response Chris.
The choice of system logic seems peculiar given the overhead that it creates if every face needs to be checked on iterating through a large collection of faces, for example when importing geometry from other file formats.

A face created with points running in a counter-clockwise direction in a plane should create a face with the normal pointing in the Z direction of the plane.
This seems to work in most cases but sometimes the entities_addface function appears to rearrange the vertexes as the code samples below illustrate. Can anybody explain why this is happening?

'==================================== A
def makeface_counterclockwise __model = Sketchup.active_model __entities = __model.entities pts=[] pts[0]=[0,0,0] pts[1]=[50,0,0] pts[2]=[50,50,0] face = __entities.add_face pts puts "--------------face.vertices" face.vertices.each {|vx| puts vx.position} end

This code returns vertexes that have been re-arranged resulting in incorrect orientation :
--------------face.vertices
(50", 0", 0")
(0", 0", 0")
(50", 50", 0")

'==================================== B
def makeface_clockwise __model = Sketchup.active_model __entities = __model.entities pts=[] pts[0]=[0,0,0] pts[1]=[-50,0,0] pts[2]=[-50,50,0] face = __entities.add_face pts puts "--------------face.vertices" face.vertices.each {|vx| puts vx.position} end
This code return expected results:
--------------face.vertices
(0", 0", 0")
(-50", 0", 0")
(-50", 50", 0")

'==================================== C
def makeface_counterclockwise_vert __model = Sketchup.active_model __entities = __model.entities pts=[] pts[0]=[0,0,0] pts[1]=[0,0,50] pts[2]=[0,50,50] face = __entities.add_face pts puts "--------------face.vertices" face.vertices.each {|vx| puts vx.position} end

This also returns expected results but contradicts the behaviour in A:
--------------face.vertices
(0", 0", 0")
(0", 0", 50")
(0", 50", 50")

As Collada is XML there is no reason why they should not be able to parse block and group hierarchies on import.
As said Sketchup exports the hierarchies OK but cant do a 'round trip' and import the geometry in the same state.

I think your comment refers to the Autodesk kit for DWG/DXF. I'm under the impression Sketchup uses libraries provided by the Open Design Alliance.

I started the evening thinking I was going to start working on a Collada exporter for my app.
As things turn out it appears that while the Sketchup Collada exporter exports the geometry complete with group hierarchy the Sketchup importer explodes everything inside the super-group on import - which kinda makes it pretty useless.

If Google is expecting us to consider Collada as a serious alternative to DWG/DXF they need to get their act together.