Jump to content

RobertHarrington

Member
  • Posts

    9
  • Joined

  • Last visited

Posts posted by RobertHarrington

  1. I take it then, Katie, that the answer is "NO".

    Working with roof surfaces, as opposed to a solid

    roof, is often very inconvenient. The method you descibed above works well only if the roof penetration is very simple. But for slightly more complicated roof penetrations, it doesn't help. For example:

    Say a skylight penetrates the roof with splayed reveals -- the hole is 2' x 2' at the roof top

    surface, 3'8" x 4' at the ceiling below, the

    reveals at varying angles.

    If the roof is a "solid", you just build the

    skylight "hole" and subtract it from the roof.

    Done.

    But... if you have to deal with a roof "surface",

    then you have to cut a hole in the top roof surface, a different hole in the bottom roof surface, and then draw 3D polygons to build

    up each reveal. That's a lot more work.

    So... I repeat ... is there anyway to convert

    a roof object to a 3D solid? For that matter,

    is there some way in general to convert a closed set of surfaces to a 3D solid (similar to the way that lines/curves can be converted to surfaces)?

  2. In VW 10 I often have a roof that the roof tool

    comes VERY CLOSE to being able to do -- but not

    quite. There's always an odd angled edge that

    needs to be clipped off, or an elliptical hole

    that needs to be punched through.

    If the "roof" which the roof tool produces were

    a solid instead of a ( PIO ' d ? ) group of surfaces, it would be easy to modify the roof

    with boolean commands ( "subtract solids" , etc.)

    But -- I can't figure out any way to convert a

    "roof" to a solid. So I end up drawing a lot

    of roofs out the long way -- which is a pain -- and which is what the roof tool was meant to avoid.

    Is there some way to convert a VW "roof-tool-produced roof" to a solid?

  3. It is usual practice when doing least squares adjustments of surveyed coordinate points to attempt to keep all the coordinates positive and

    in the same quadrant. (Upper right, "North East"). Tis also common to run a traverse around

    a section of land, one mile square. In this case, my control point was the Northeast corner of a section. A section is roughly 1 mile on a side, 5280 feet more or less. (But only more or less). In such a case I typically assign the Northeast corner the arbitrary coordinate of (10,000.00, 10,000.00), thus assuring that all coordinates of points within the section are positive. I'm required by statute to run

    traverses to accuracies of better than 1 in 20,000, and typically by contract to accuracy of

    better than 1 in 100,000.

    My layer scale is 1:2400, 1" = 200'.

    The above is nothing unusual -- a random land surveyor on any given day will run traverses

    much larger and more accurate.

    If the layer scale actually has an influence ( I

    can't see why it would, but if it DOES.... well)

    I suppose that for COGO purposes, it wouldn't bother me to set it to 1:1 (although I'm not sure

    if VW would choke on the required paper size -- not

    that I'm planning on printing anything out at that scale !! )

  4. I'm on a Dell PC (P4, 2.4 Ghz, 768 Meg RAM)

    and using feet. The initial coordinates of the

    point are, oh say, x = 8,500.00 and y = 10,000.00. Then go 30.00' from there.

    I will try 10.5.1 if/when I get over various

    time critical projects.

    An added point -- it would be nice if VW allowed

    angular accuracy to be input/displayed consistently in units more accurate than seconds

    of arc -- say hundredths or thousandths of a second at least.

  5. RE: VW 10.1.0, Win XP

    I have used AUTOCAD to do 2D coordinate geometry (COGO) for land surveying. I HAD hoped to do the same with VW. But... VW does something bad, approximate and irritating that makes it useless for COGO. It makes terrible round-off errors.

    I've got all snaps off, linear and angular "units" set to maximum accuracy (lotsa

    zeros). I've set angular units to "degrees, minutes and seconds.

    Say I want to draw a line of a certain length, at

    a certain angle (a very basic thing to do in

    land surveying, it's called a one course of a

    traverse -- a traverse typically containing

    dozens of courses)

    Hit the line tool. Click the mouse anywhere on

    the screen. The beginning point of a line is set.

    Now set the angle:

    Tab over to the angle entry box box on the data display bar. Enter "-89d59m26s" in the angle entry box. Hit return. Angle Data is accepted

    correctly.

    Now to set the length:

    Tab over to the "L:" (length entry) box on the

    data display bar. Enter "30.00000" Thus far, so good. Hit the return key. DATA IS NOT ACCEPTED CORRECTLY !!

    Upon hitting the return key, the "30.00000" is CHANGED by VW to "30.0117235714" !!! (This also

    is what shows up in the object info palette).

    Now, look, I just SPECIFIED 30.00000 -- what gives with VW throwing in an additional 0.0117235714 ? A rounding error of 1 in 3000 is

    NOT GOOD !!!!

    Interestingly, if I key in the same angle, but

    change the length from 30.00000 to 3000.000 --

    VW STILL changes the length I entered from 3000 to 3000.000177649 - an acceptable degree of rounding error 1 part in 30 million or so.

    I also note that, depending upon the initial point of the line, VW gives me back slightly

    different answers.

    I also note that no matter what I do up in the

    data display bar, I can't keep VW from CHANGING

    the entered data and I can't get it to leave

    my length of 30.00000 ALONE. Plus, while I can't

    get the data display bar to accept 30.00000, I

    CAN, once the object is entered, go to the object

    info palette, set how it displays a line object

    info to "angle and length" and change the funky 30.0117235714 BACK to 30.00000.

    What is going on? Is there ANOTHER "units" or round-off parameter to VW I can set to get more

    numerical accuracy?

  6. That was absolutely it, Kevin. Should have

    thought of it myself, but I got my mind so

    wrapped around the problem that I missed the obvious.

    Thanks a million. As I recall, that was the

    cure to someone else's problem 18 months ago.

    It's just that the Tech Board Boolean search

    won't turn it up when you are thinking the

    search terms are "PIO or symbol or selection"

    Once every 100 instances or so, I mean to change

    "Organize>Layer Options>etc." and accidentally

    change "Organize>Class Options>etc." instead.

    Then, if I don't notice it at the time, it comes

    back later, unexpectedly, and drives me NUTS.

  7. VW 10.1.0, Windows XP

    I've done something odd to a drawing. About

    half the hybrid 2D/3D symbols I create "get lost". That is, I insert them from the resource

    browser, (and there they are) and they are viewable in 2D and 3D, but they are no longer selectable and, more painfully, no longer eraseable.

    And the REALLY FUNNY thing is -- it happens to

    the same symbol in the same drawing. Right now

    I'm looking at a drawing with 2 chests of drawers -- both symbols, both the SAME symbol.

    One is selectable, one not.

    The only way I can get the non-selectable symbols

    to go away is to delete the layer.

    I have some recollection of Katie addressing this

    about 1.5 years ago, but I couldn't find the post.

    Any suggestions to cure my (probably self-induced) problem?

  8. I've long used VW (MiniCAD) for 2D drawing (mainly 8.5.1). VW's "ease of use" has been one of its biggest selling points. However,

    I've always used other programs for 3D.

    Recently I spent an hour with the VW10 demo,

    hoping that VW's 2D "ease of use" would carry

    over into 3D. My conclusion is -- while the 3D

    tools are pretty capable -- the lack of "viewports" makes VW10 unduly difficult to use.

    For example, in AutoDesk's (5 year old) 3D Studio Viz -- one can simultaneously view orthogonal

    top, bottom, side AND perspective views (or any

    4 sets of random working planes). This makes drawing, editing, etc. vastly easier -- one can draw in a convenient orthogonal view (or VIEWS) while observing the result in all other views. One doesn't, for example, inadvertently drag, drop and create a cuboid BELOW the ground plane, when one meant to draw it ABOVE same.

    Did I miss it during my demo tour, or is VW10

    limited to ONE view at any given time? And, if

    so -- any chance for multiple viewing windows

    in a future version of VW?

×
×
  • Create New...