Jump to content
  • 0

Real-time Elevations


Ken

Question

I'm a long time user of Vectorworks and take advantage of every feature available (except very lengthy Vectorscripting). I use highly customized template drawings and an ever-growing library of hybrid symbols for all my projects.

In my typical architectural project, I create, inside one file, all existing drawings (the "as-built" plans, model, elevations, site plan) as required, all proposed drawings (scheme alternates, plans, models, elevations, etc.), revise and develop SEVERAL 3D models, export to Art?lantis for rendering as required, develop structural foundation and framing plans and sections, show the full Site Development Plan at a different scale (if not extremely complex), then develop everything into a FULL SET of highly detailed construction drawings with imported PICTs and BITMAPs as required. Worksheets compute areas and quantities as required. I have a lengthy procedure that exports to Autocad for consultants' backgrounds if required. Everything except large-scale detail drawings and specifications are in ONE electronic file. My projects are mostly residences (up to sixplex townhomes), but most small building projects fit this approach. New construction in VW is even easier than renovations.

Having said all that, I'm posting these wish items that I think would take VW to the next level -- automatic and REAL-TIME architectural elevations. It seems a lot of programming effort was made in creating factory add-ons like VA to help the beginner. I would never use VA. That is NOT the next level. As a loyal and seasoned VW user I'm concerned about the direction that VW is taking -- the feature-packed direction instead of creating stable, quirk-free and bug-free innovations.

For real-time architectural elevations my wish items include:

1) Finally fixing walljoins for Y and complex T intersections! Make them seamless in 3D as well as 2D -- for model views, top/plan printing and DXF export.

2) Create a way to assign DIFFERENT LINEWEIGHTS for layerlinked objects for each layerscale where they appear.

3) Create a way to assign different lineweights for SYMBOLS for each layerscale where they appear.

4) Create more snap points in extrudes, 3D polys, walls, roofs, etc., in straight and iso views.

5) Reveal more snap points in layer linked objects.

6) Create a way to layerlink a layer link -- at least two generations. Alternately, create multiple view ports that can be made into a *printable* elevations sheet.

7) Allow dimensioning of a 3D model in front, back and side views.

Can you guess what all these improvements will add up to?

Does VW9 offer these improvements? I do not have VW9, but it's doubtful from what I've seen and read. I believe these improvements would add up to quite an innovative feature. Vectorworks would gain the "killer feature" advantage over other 2D/3D CADD of REAL-TIME elevations that need not be touched again after designing in planview and for the model -- suitable for either presentation or construction drawing prints.

[This message has been edited by Archken (edited 08-19-2001).]

Link to comment

Recommended Posts

  • 0

Archken,

I found a solution for the printable elevation sheets. Link your model together in your elevatio sheet and then duplicate the link 4 times (or whatever) and rotate each link to display a different elevation. Even if you duplicate and reposition a layer link it will still update in real time.

Good luck

Link to comment
  • 0

MikeB, that is precisely the problem! It is unfortunately the best that VW8.5.2 can do for now. I doubt VW9 has any improvements for elevations, except perhaps Extrude Along Path, which definitely helps when creating trim.

The problems (again) are:

1) Wall seams at Y and complex T walljoins are revealed in the model views,

2) Lineweights are same as planview (my floor plan walls are appropriately HEAVY),

3) Window symbols have lineweight problems for trim and openings in model view,

4) Stretching and snapping walls/ extrudes/ 3D polys are still abysmally tedious (not to mention that edit extrude no longer aligns in plan view like we had in Minicad7!),

5) Difficulty snapping to linked objects

6) Re-linking for each view is required (four times the effort as one model)

7) Dimensions must come off 2D objects

So my wishes above (please READ) are targeted at each of these problems. Then, and only then, can we have true REAL-TIME elevations -- something worthy of a new version of the software.

Link to comment
  • 0

Sorry Archiken, I spoke before I thought

I've used this technique on renderings done with RW. Our jobs are too big, and usually too complicated to try to use the wall and door 3-D info. We typically have seperate files for each sheet in a set of drawings so different people can work on different sheets at the same time. Consequently we have to draw our elevations in 2-D. I hadn't looked at the line weight problem because I'd only used the Raymaker rendering engine.

Link to comment
  • 0

MikeP,

Too often I've seen, and worked on, large architectural projects where VW is used purely as a 2D drafting tool. Then somebody on the team gets the revelation that VW can be used for *modeling* as well. Some parts or phases of the project then get the VW "model" treatment, perhaps exporting to fancy rendering software for photo-realistic output The rest of the project continues line-for-line, groups-and-groups ad nauseum tediousity, get-my-resume-out-for-a-better-position-elsewhere drafting and confusion. I've been there.

For architectural production management nothing else may be suitable. However, it does not support evolution of VW into something better than we had in Minicad 6!

In my opinion, VW is *almost* there. It's not Archicad, which comes in from the top with 3D modeling and has limited drafting capabilities. It's not Autocad, which proliferates with a "2D line segment" mentality and builds with add-ons after add-ons, creating an ever-challenging entrance into the College of Autocad. DenebaCAD is not quite there yet. Allplan sits in another market altogether and too expensive.

I hate to think I'm stuck with Vectorworks the way it is. I think it only needs to address some fundamental quirks like Y intersection walljoins to feel "next generation." NNA, please tell me VW9 or some subsequent part of it will address this.

Link to comment
  • 0

NNA, I think you should pay for two days of Archken's time to sit down with you and show you, with his own projects as examples, exactly what he is talking about. I don't doubt that it would be difficult to implement even the simplest-sounding aspects, but I am intrigued.

With respect to version 9, I have been working with 9.0.1 (Windows) for a few weeks, mainly to do some work to illustrate an academic paper. My feeling is that it is much more stable and bug-free than the comments on the bulletin board would make it seem, although I haven't used many of the production-drawing tools yet. For one thing alone, Archken, you would find it worthwhile: you can assign many more keyboard shortcuts since you now have <Ctrl + shift + >, <Ctrl + Alt + >, and <Ctrl + Alt + shift + > options. Also, the tool palettes dock to the top or side, which makes your workspace much less cluttered. You may use RenderWorks or some other 3rd party rendering program, but the Open GL rendering mode that comes with VW 9 is a great improvement, both in terms of allowing a range of options and much better interpolation. There are a bunch of other small things that I'm not mentioning.

BTW, two minor picky points: setting up 4 linked layers to show different views of the model doesn't seem to be too much of a burden! Also, the use of the term "realtime" - I think "automated" is what you're thinking. "Realtime" refers to the ability of a computing system to perform calculations fast enough to keep up with real events, as opposed to it having to spend 20 hours, for example, to create a video that takes 2 minutes to view.

VW 9.0.1

Win 2000

PII 333

Link to comment
  • 0

just to throw my 2 cents in here - Archken is dead on right. I would like to suggest to NNA that they sit down with a copy of Revit and take a good hard look at what that program offers. I would be working on it right now, if only Revit was available on the Mac. That program actually uses 3d for everything! you can cut a dynamic section (sim to archicad) modify elements of the building and have it update in real time in every plan, section, elevation, schedule and so forth. and it does not stop there.

granted revit costs a bit more that VW, but then again minicad cost me less than $500 when I first started using it and now look at it. I would gladly pay a bit more to have something that utilized a fully dynamic model and did not force you to create static snapshots.

i guess that is a bit more than 2 cents, but now i feel better!

Link to comment
  • 0
Originally posted by Archken:

"...It seems a lot of programming effort was wasted in creating factory add-ons like VA to help the beginner. I would never use VA. That is NOT the next level. As a loyal and seasoned VW user I'm concerned about the misdirection that VW is taking -- the "featuritus" competitive approach instead of creating stable, quirk-free and bug-free innovations."

1 - I like this general point and want to support it, based on my experience with VW. I've purchased VA and intend to learn it, try it, and work with it. By its nature though VA (versus VW) seems to consist of many many single purpose tools, each one requiring its own learning curve. From my perspective as a sole practitioner architect, working sporadically with the program, this isn't ideal. That the core tools (VW) are effective and widely useful seems far more important. As the cumulative required learning curve increases, the attractiveness of the program to general users (people who actually draw on the computer maybe 25% of the time) diminishes.

2 - My experience: "Draw it once" may be fine as a sales' approach, but as a literal plan for design, not workable. Again, my experience, but the design process seeems to be not a straightforward accumulation of detail in a drawing...rather the inevitable drawing and redrawing of plans, elevations, sections to work out ideas. My point is that the plan-based 3d drawing process in VW doesn't yet work for this architect, and I find that I have to construct / design elevations and sections as well as plans. And not only once, but repeatedly. The virtues of CAD and paper/pencil are both considerable in this process. Perhaps "realtime elevations/sections" would, if they could, help address this concern. Don't know.

Link to comment
  • 0

Well, yes. My terminology may be incorrect. "Realtime" refers to something a bit different. We already have realtime views in a linked model.

"Automated" is indeed closer to my VW wishes. Ideally, of course, we would simply speak into a magic box, "Computer, produce all elevation drawings!" That will be the day! In present reality, however, I still think VW is *almost* there.

My simple wish for the ability to design and compose drawings just ONCE into usable, printable sheets, then have subsequent design modifications update themselves automatically in every drawing is indeed closer than we think. We can, in fact, do this in 8.5.2 for all planview drawings from enlarged floor plans to building plans to structural framing plans to mech/elec plans to site plans -- I can and I do -- so how about elevations now? If only 4 of my 7 wishes are granted, I can prove how every architect should love VW! NNA, I'm really trying to be your evangelist!

Regarding relinking, I must disagree on how burdensome it really is for the four or six side views (e.g., U-shaped plans). The easiest way, actually, is to unlock and duplicate the whole model into "elevation view" positions on a sheet. But it's still rather unwieldy in the X,Y,Z directions for "heavy" models with several light sources. Layerlinks are again not very snappable/grabbable. A separate linked model is still needed as a "study" model. Every subsequent change in layer Z then requires re-linking, re-unlocking, re-duplicating, re-positioning, etc. (perhaps a quirk to be fixed?). Furthermore, when 3D elements are placed only in the model layer (e.g., end conditions in multi-family attached residential), it gets all the more unwieldy. Then there's no quick way to detect how many linked objects are in a particular layer (VW8.5.2). If only a bunch of stuff can be marquee-selected and have the OI palette list them by their type and source layer... geesh, that's another one of my old old wish list items!

In general, my wish to layerlink another layerlink comes from situations where layerlinked objects are used only as *components* of a layer that is a source layer in itself. I think a maximum of two generations would be useful. LLs are so powerful, why not expand just a little?

With MacOS X knocking and several major design projects on my horizon, I'm ready to upgrade everything -- even my primary CADD software. Have I outgrown Vectorworks? I hope not. Do NNA people read these (their own) pages? I hope so. In the meantime, I will definitely look into Revit -- perhaps to run in VirtualPC?

Nice to see positive remarks on VW9.0.1.

Link to comment
  • 0

Here are suggestions for taking baby steps toward developing automatic realtime elevations:

1) Floor-to-Floor Seams: Create a way to automatically hide them when two floor layers are linked into a model. Currently the least number of seams is ONE: when the top edge of a lower level wall meets the bottom edge of an upper level wall. Alternatively, create an option to turn this seam into a 0.03mm (1 mil) line. This is acceptable because it would act as the "floor level" reference line on the printable elevation drawing (not shown below).

2) Layerlinked Lineweight: Create a way to change lineweight for each linked object. The model view below shows lineweight carried over from the floor plan, where walls have appropriately heavy lineweight. Printable elevations need heavy lines for foreground "cuts" and for the "profile" of the entire building -- not at every corner.

seam1.gif

[This message has been edited by Archken (edited 08-24-2001).]

Link to comment
  • 0

3) Vertical Wall Seams: Create a way to hide vertical wall seams where walls of different thicknesses are simply butted up against each other. Somebody a long time ago (on MC7) suggested adapting the "hide edge" feature from 2D polygon edit for 3D walls. This may work for both vertical and horizontal seams.

4) Intersecting Edges: Here's where we want the seams to show. Currently (8.5.2) there are no edges in hidden line rendered view at the intersection between a roof, wall, extrude, slab or mesh.

seam2.gif

[This message has been edited by Archken (edited 08-24-2001).]

Link to comment
  • 0

5) Round Object Segments: Create a way to hide all edges of round walls in Hidden Line render mode. Round extrusions also need a way to hide every edge except profile. I've been using octogons instead of circles for extrusions when 45? side views are necessary (like southeast in addition to south and east) -- but that still leaves two edges to delete. Sweep objects are same.

seam3.gif

Link to comment
  • 0
  • Vectorworks, Inc Employee

quote:

Originally posted by Archken:

5)
Round Object Segments
: Create a way to hide all edges of round walls in Hidden Line render mode.
seam3.gif

Your wish was my command. In VW9, a simple VectorScript of SetPrefReal(72, angle);

where angle is somewhere between 0 and 90 degrees, will get rid of a whole lotta facet lines. Experiment to figure out the best value, maybe 20 degrees?

Sorry this didn't make it in the standard UI, we just ran out of time.

------------------

Andrew Bell

andrewb@nemetschek.net

I am not an official spokesperson for NNA

Link to comment
  • 0

Andrew,

Good work! Actually, the preference should work the other way around. We hardly ever want the extra lines, and almost always want things to look the way Archken suggests. Go Archken!

The lack of lines showing the intersection of solids, as per his earlier suggestion, is a constant problem. It is possible to get them to show up if you perform an actual boolean solid addition - but that doesn't work with walls and roofs (?), and not very convenient with the solid modeling of other kinds of objects - especially now that we have the accurate but slow VW 9.

Link to comment
  • 0
  • Vectorworks, Inc Employee

The defaceting code was finished late enough in the development of VW 9 that it didn't get the polish it could use. Really, I wanted a checkbox for defaceting all known curved surfaces (on by default), with the angle only relevant for meshes. Maybe by VW 10, it'll get its own "hidden line preferences" dialog, like OpenGL and RenderWorks rendering have.

Walls and roofs do work as solids, although if you have different textures for the different faces that gets lost when it becomes a solid, and the special texturing for roofs disappears.

Are you working on a Mac? I've been doing a decent-sized project with several solids on a PC, and while solid computation isn't instantaneous, the speed wasn't slow enough to be an issue. I've even considered turning sweeps into solids so there aren't so many lines in wireframe (by subtracting objects that don't intersect with them).

------------------

Andrew Bell

andrewb@nemetschek.net

I am not an official spokesperson for NNA

[This message has been edited by Andrew Bell@NNA (edited 08-23-2001).]

Link to comment
  • 0

Thanks, Andrew "The Genie" Bell! I'm looking forward to that script (currently I don't have VW9 to understand it... hey, that's a clever marketing tactic!).

Is it anything like SetPrefInt(56,integer) in 8? I'm assuming that Hidden Line render is still the best way to view models as elevations, and that it will still be available in 9. Would SetPrefReal(72,0) produce one smooth curve with no visible edges in each of four side views? That would be absolutely fabulous!

seam5.gif

Link to comment
  • 0

6) Complex Wall Intersections: Eliminate seams at complex Y, T and X wall intersections. Top/plan walljoining is, of course, the highest priority (hoping for a thorough solution that encompasses ungroup to clean lines for DXF export). Corresponding side view seamlessness (hidden line render) is then required to get nice elevations from it.

seam6.gif

The more I look at this problem, the more containable it seems as far as programming a solution -- from my untrained end-user's point of view. There are only a handful of situations that arise over and over again, so it's not an infinite set of possibilities. The Y intersection shown is one of perhaps only two truly awkward wyes. The T intersection is one of perhaps two or three situations, with the thickness of the adjacent walls being key. The X intersection is similar.

Link to comment
  • 0

7) Cut 3D Section: I don't understand how useful this is (VW8.5.2). Wall symbols slice the wall into separate 3D polys, causing edge lines to show. It requires Converting To Lines (as opposed to keeping in Hidden Line render), which creates a tremendous amount of 2D editing. A building section drawing may need to show the uncut portion beyond in elevation view. This needs some kind of fix.

seam7.gif

Link to comment
  • 0

quote:

SetPrefReal(72, 0); disables the facet line removal...

Then how about SetPrefReal(72,1) -- would that create one smooth curve looking from all four sides (front, rear, right, left) in Hidden Line render for the example shown?

seam5.gif

Link to comment
  • 0
  • Vectorworks, Inc Employee

quote:

Originally posted by Archken:

Then how about SetPrefReal(72,1) -- would that create one smooth curve looking from all four sides (front, rear, right, left) in Hidden Line render for the example shown?

No, more like SetPrefReal(72, 20), as the faces of the round wall are about 15 degrees rotated from each other. The parameter is a value in degrees: 0 would be coplanar, 90 would be a square corner. You want a parameter that is larger than the angle between faces that you want the lines removed. It takes a little experimenting to get the value you want.

------------------

Andrew Bell

andrewb@nemetschek.net

I am not an official spokesperson for NNA

Link to comment
  • 0

quote:

Originally posted by Andrew Bell@NNA:

Are you working on a Mac? I've been doing a decent-sized project with several solids on a PC, and while solid computation isn't instantaneous, the speed wasn't slow enough to be an issue

Andrew:

I have solid booleans that take 30 seconds or more to compute in VW 9.0.1 (timed at 8 seconds in VW 8.5.2). If we have to do this kind of thing to get lines showing the interesection of all solids, it's too slow. Consider that you will have to go in and edit all of these complex objects every time you want to change one, and wait each time for the calculation to be completed.

Thanks for your participation in this forum!

-Pete

VW 9.0.1

Win 2000

PII 333

192 MB RAM

Link to comment
  • 0

The final baby steps , in my opinion, toward creating "automatic" elevation drawings are as follows:

8) Symbol Lineweight: Create a way to change a symbol's lineweight for different layerscale placement. For example, a layerscale half as large requires symbols' lineweights half as thick. This will actually help planview drawings of a different scale more than elevations. Wall symbols are most important -- currently their lineweights become extremely tedious to adjust for different scale elevations with complex windows.

9) Plug-in Objects: These are excellent beginnings of something great, but they are still very limited by: a) cannot set to wall depth when made into symbol, b) 2D view cannot be kept simple and different from 3D version, c) some parts cannot be adjusted, like mullion thickness, sill thickness, apron, stool, slope d) how can a different lineweight be assigned to each part? How do you change the mullion and glazing "styles?" I hardly use it. I don't know.

10) Hatches and Textures: I've been drawing 2D polygons for hatches to show materials in elevation view like roof finish, stone, concrete. Can 3D textures be made to appear decent in Hidden Line rendered view? (I don't have this feature so I have absolutely no idea if it does already or not, but I suspect not)

VW8.5.2/mac

Link to comment
  • 0

11) Printed Output: I hope NNA software programmers are aware that architects need to print their ideas and drawings to make money. Rarely do we get paid from on-screen digital presentations alone (assuming some are even doing THAT). As a result, printed drawings are pivotal to success. Conversely, lack of good prints spells failure.

So what type of printed output count the most?

In general architecture, three categories with time-effort averages are often cited:

1) presentation drawings (schematic design): 15% of the project

2) development drawings: 20%

3) construction drawings: 40%

4) non-drawing tasks: 25%

For small structures, development and construction drawings are often rolled into one phase. They may look something like this:

BBrender.jpgBBelevs.gif

An average of 60% of project time (often more for larger structures) has nothing to do with rendering, but everything to do with flat production drawings. As a long-time VW user, I've taken advantage of 3D as much as possible to speed up my production drawings. Still, elevation drawings require a lot of 2D editing -- too much unnecessary 2D editing, in my opinion.

Appropriately, Vectorworks (and previously Minicad) has established a solid market position with superior drafting features. Bravo for VW. This is why we all love it. Archicad, on the other hand, has always lacked 2D drafting power. Autocad, as another example, is not so cost-effective because it's so time-consuming to use competently. You have to go an Autocad college to learn drafting! In my opinion, VW is a front-runner by having both 3D and drafting features.

So the big question is: what does an "integrated 2D/3D Vectorworks" really mean for 60% for our work, which rely on printed drawings?

This question precisely becomes: Must VW users go back to line-per-line 2D editing after a 3D model is created?

Or do we have two different camps of VW users -- the 2D camp and the 3D camp?

How does VW9 or VW-A improve on previous versions in printed output? I know Extrude Along Path will help. There is still, however, in my opinion, a significant gap in 2D/3D integration, which has remained unchanged since MC6.

I'm simply suggesting small improvements that will add up to quite a punch when all put together: more integrated 2D/3D production speed and accuracy! Automatic elevations can indeed be better than we have now in 8.5.2 (presumably same as 9). I'm wishing for the next generation of architectural elevation drawings that just need notes, dims and perhaps a heavy profile line drawn on top of Hidden Line render to be presentable.

[This message has been edited by Archken (edited 08-30-2001).]

Link to comment
  • 0

Archken,

I am adding to the thread you started some time ago because I find myself in the same position of having maxed out on VW, wanting to go to the next level of live-linked elevations and sections. While versions 9 and 10 contain worthwhile improvements and I wouldn't hesitate to upgrade, Revit seems to be a leap to another generation of software. Could you update us on your current thoughts on this?

Link to comment

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Answer this question...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...