Urgent Worry

After the March 8, 1987 programmers' meeting which launched the 3D effort which ultimately produced Release 10 (see page ), work continued on Release 9, aiming for shipment as soon as possible. Upon his return from the first AutoCAD Expo Europe in Birmingham England, where everybody seemed to be asking us for more 3D, Kern Sibbald wrote this memo to propose adding additional 3D capability to Release 9, even if it meant slipping the schedule to do so. His proposal, called “Urgent Worry” (a play on my “Urgent Fury” project which formed a main theme of Release 9), was ignored; Release 9 contained almost no additional 3D functionality.

This significance of this memo lies in its spelling out, in a precise manner, what new features would be required to create a genuine 3D AutoCAD. Kern's specifications here formed the kernel of the 3D feature set implemented in Release 10. In retrospect, we would have been wiser to heed Kern's recommendation of merging the Release 9 and 3D efforts; customers and dealers objected when we shipped Release 9 only 5 months after Version 2.6—they prefer updates no more frequently than once a year. When we shipped Release 10, many users reacted, “thanks for all the 3D stuff, which since I do 2D drafting is totally useless to me,” and we had to play catchup on drafting functionality in Release 11. A merged Release 9 and 10, shipped in late 1988 would have been much better received.

Urgent Worry

by Kern Sibbald — April 10, 1987

We have heard a lot about 3D AutoCAD lately. In fact it has become the number one technical priority of the company. This is with good reason, because our users want 3D. Currently, it is probably the most asked for feature in AutoCAD. AutoCAD 2.6 with 3D level 2 will partially satisfy the requests for 3D. My impression from the shows is the users expected more, but they are happy to get some 3D features and are willing to wait a while for more features, but not very long. All of this you have probably heard before and fits with our current strategy of accelerating the 3D project, which I fully support.

At the Birmingham conference in England during the many passionate pleas for additional 3D features, it occurred to me that with AutoCAD 2.6 3D level 2, we have made it clear to our users that we are going to support 3D. And since users clearly want more 3D features and are expecting them, it also occurred to me that if we release an AutoCAD subsequent to 2.6 (White Album) with no new 3D features, we will unwittingly send our users a strong message saying that we not really serious about adding 3D to AutoCAD. Our users want more 3D now, but are willing to wait a little longer. I believe that if we add no new 3D features to White Album, we will seriously erode our what remains of our credibility for creating a 3D product. This could be a very serious mistake. Place yourself in dealer's shoes for a moment and think about what this would mean. On the other hand, if we show progress toward a full 3D by adding some new features, we may be able to retain a neutral position in our dealers and users eyes.

What would the impact on White Album be adding new 3D features at this time? I don't know the answer to that question. But I do feel that it would not seriously detract from the “real” 3D effort if we spend about 4 hours designing a few of the features that we want in the “real” 3D. After that, we can answer the question.

Below I have listed a few of the features that we should consider for immediate implementation in White Album. All these features can be added in a very short time (I estimate one week) if we form a small team of programmers (no more that 5) who will work together to add these features. The time seems right to do it since 2.6 is out the door, and 2.7 has been dropped. My understanding is that although work has begun on the “real” 3D, none of the 2.7 programmers are yet working on it. This would seem to leave a number of programmers itching to do a quick kill project.

Possible New AutoCAD Features:

• Add a new 3DVIEW dialogue box—or enhance VPOINT to include parameters necessary to add perspective and to adjust the view point:

  polar rotation and twist
  look at
  look from
  lens
  film diagonal
  perspective
  view distance

• Make every command accept a 3D point as the first input. If the Z value is subsequently changed (in a 2D command), print a warning message and disregard the Z component.
• Create a “primitive” CPLANE command that allows the user to toggle from the current X-Y plane to the X-Z plane, and Y-Z planes for cursor/keyboard input. Similar to ISO toggle.
• Implement a 3D rotate command that drags 3D entities and allows use of the CPLANE toggle.
• Add a Z coordinate to the status line.
• Add additional object snap modes in 3D-mid of 3DFACE, center of 3DFACE, perp of 3DFACE, perp to 3DLINE?
• Implement a Z grid and Z snap that will be activated when the “CPLANE” includes the Z direction.
• Eliminate interior faces from polylines, extruded circles, etc.
• Mode to make the LINE command behave as the 3DLINE command does.
• Make new 3D primitives for BOX, SPHERE, CYLINDER, CONE, DISH. Allow them to be generated about any arbitrary axis.
• Add some way to group or block the 3DFACEs in BOX, SPHERE, etc.
• Add a volume clipping dialogue box as in Shade. Normally the volume clipping will be automatically established when perspective is invoked according to the film diagonal specified. However, the user may wish to explicitly control clipping—for example when producing section slices of a drawing.
• Explode a polyline into 3DFACEs (possibly a mode).

Future Features:

• Add a numeric box. This box could be in the lower right corner of the menu and would contain a numeric keypad with some special characters like backspace, +, −, and return. It would permit the user to enter any number with the use of a mouse or tablet.
• Implement a configurable status line using C printf format and variables names as in a printf command. “%d %s …”,COLOR, LAYER, …. Feed the names into bag as if we are Lisp asking for variables.
• Need a program that will verify that the sizes of dialogue boxes are correct—mostly for the foreign language versions. Could also implement variable size dialogue boxes similar to the alert boxes.