Revenge of the paper space cadets

In my July 2000 column, "Lost in paper space", I took on the contentious question of whether to use paper space as your standard operating procedure to set up all AutoCAD drawings. I compared several xref and paper space methods and argued that a model file and sheet file approach using only xrefs and model space solves many of the presentation problems that paper space is designed to address. I argued further that the xref-only methods are in many cases simpler, more flexible, and more robust.

That article unleashed a flood of e-mail from both supporters (the x-ruffians) and dissenters (the paper space cadets). This month, I’ll pass on and respond to some reader comments and also cover some additional techniques and ideas.

Can’t we all just get along?
To reiterate, I was not arguing that paper space is useless. It’s indispensable for some kinds of work—notably presentation of 3D models—and useful for others. I argue against an unexamined acceptance of paper space as standard operating procedure and for a reasoned examination of the merits and disadvantages of different paper space and xref methods. All AutoCAD users should be comfortable with a variety of drawing organization methods, including ones that use paper space. This familiarity helps you choose the methods that work best for your industry, office, and project. It also ensures you’ll be able to understand and work with the different kinds of drawings that your clients and coworkers throw at you.

Your responses reminded me that different CAD industries have different customs and requirements. AutoCAD users in the mechanical industry are almost unanimously in favor of paper space as standard operating procedure, while those from architectural and engineering offices are more divided. Most of my experience is in creating construction documents for buildings, so my advice and opinions invariably reflect that perspective. In particular, the model file and sheet file approach to creating plot sheets appears to be most applicable to building drafting.

I recently helped two client firms with major revisions of their CAD standards. Both are structural engineering firms, and they have roughly the same number of full-time CAD drafters who use AutoCAD and engineers who review and plot drawings using AutoCAD LT. In each case, we considered the "how to organize plot sheets" question early on and debated the merits of various paper space, xref, and hybrid schemes. Interestingly, the two firms came up with very different decisions about AutoCAD drawing organization. One decided on an approach that uses paper space for all plot sheet drawings and limits xrefs to background drawings— for example, an architectural floor plan that serves as the background for the structural framing plan. The other firm chose an approach that uses model files Xrefed into a sheet file (all in model space) to create plot sheets.

Both firms made good decisions. The differences were due in part to office customs that had developed around each office’s old CAD standards and in part from honest differences of opinion about which advantages and disadvantages counted more. The moral is that there’s no one right way to do things. Make sure that the decision-makers understand what they’re talking about (that is, that a preference isn’t due to ignorance), get all of the opinions on the table, hash them out, be willing to compromise, and agree to abide by the final decision.

Advantage: paper space
Several readers wrote to point out some particular situations that I didn’t mention where paper space holds an advantage over other methods:

• If you must frequently adjust the size of a clipped view, paper space viewports are much easier to edit than xref clipping boundaries. With a paper space viewport, you simply make sure that the viewport’s layer is on, click on the viewport border, and grip-edit it. With a clipped xref, you must delete and recreate the clipping boundary.
• Paper space is a convenient way to present civil engineering plan and profile drawings, with the plan in one viewport and the corresponding profile in another.
• For building plans, some people use a stacked plans approach. You draw all of the 2D plans for a building on top of each other, but on different sets of layers (one set of layers per plan level). You then create a paper space viewport for each plan and turn on only the layers for the particular plan that should appear in that viewport. This method works well on smaller projects as long as you’re fastidious about putting objects on the right layers. I think that this method quickly gets out of hand on larger projects, such as multi-story office buildings. There’s just too much stuff to deal with easily in one file. Also, on larger projects, you often need to divide the work among multiple drafters, so locking everything up in one DWG file is a problem.
• You can use paper space during the CA (construction administration) phase of a project to create an addendum or clarifying sketch that supplements the construction documents. In the old manual method, a CA engineer or architect photocopies a part of the drawing onto ordinary paper, sketches changes or additions in response to a contractor’s request for information, and faxes it to the job site. With AutoCAD, you make a copy of the submittal drawing, create a paper space viewport of appropriate size, zoom and pan until the area of interest is visible at an appropriate scale, and make the changes or additions.
• If you’re willing to live with your annotations and geometry in different spaces (i.e., add text and dimensions in paper space instead of model space), you gain a certain amount of simplicity (no more messing with weird text heights) and flexibility (showing the same geometry at different scales).

Different layer settings with clipped xrefs
Even we x-ruffians have to admit the advantage of paper space viewport layer settings. You can create two or more viewports that show the same model and change layer properties separately in each. For example, you can draw a floor plan and reflected ceiling plan together in model space, then show it in one paper space viewport with the reflected ceiling plan layers visible and in another viewport with only the floor plan layers visible.

But we won’t give in without a fight. This trick lets you show layers differently in different clipped versions of the same xref. Here’s how to xref and clip a drawing called PLAN.DWG twice, with different layer settings each time:

1. Xref and clip the file once.
2. Reopen the xref dialog box and rename the xref to something different from the filename. For example, change the name from Plan to Plan_clg.
3. Xref and clip the file again.
4. Optional: Reopen the Xref dialog box and rename the xref from the previous step to something more recognizable. For example, change the name from Plan to Plan_flr.

Now you have separate sets of xref layers for each attachment, which you can manipulate separately in the layer dialog box. In this example, you have a group of layers named Plan_clg|* corresponding to the first instance of the xref, and a second set named Plan_flr|* corresponding to the second instance of the xref.

I admit that this method is less direct than putting the cursor in a paper space viewport, opening the layer dialog box, and turning off layers by clicking in the Active VP Freeze column. But the "modified xref name" kludge does have one advantage—you can change other layer properties, such as color or linetype, in individual instances of the xref.

The ultimate in paper spaceyness: detail sheets
I’ve argued vehemently and often against one particular use of paper space: assembling detail sheets that contain details of different scales. In the common version of this approach, a hapless paper space cadet creates a flock of paper space viewports, one per detail, and zooms and pans inside each one until the appropriate detail appears. After all of that thrashing about in paper space, you often end up with performance problems.

The July article describes my preferred xref and model space method for creating detail sheets, but I’ve recently run across a different paper space method that’s considerably more elegant. This new method is a bit tricky to describe and name, but let’s call it multiple, overlapping, full detail grids.

Start with a checkerboard-like grid of detail modules that fills the drawing area of your sheet. For example, you might have a 4X3 grid of 7.5"X7" detail modules, which fit the drawing area of a 36"X24" sheet with some room left for margin and title block. Make one copy of this full-detail grid in model space for each detail scale that you want to use (1/2"=1'-0", 3/4"=1'-0", 1"=1'-0", and so on). The detail grid for each scale is scaled up by the corresponding drawing scale factor (24, 16, 12, and so forth.). In paper space, you create one large viewport

Figure 1. The multiple, overlapping, full detail grids detail sheet method.

Here are some additional things to keep in mind:

• As you work on details of different scales, you must change scale-dependent settings, especially dimension scale. You’ll probably want to create a separate dimension style for each scale.
• Because all of the paper space view-ports overlap exactly, you can’t reliably move the cursor into a particular viewport by clicking in the viewport area. You can use the keyboard shortcut <Ctrl>–R to switch among the overlapping viewports.
• Before you plot, set PSLTSCALE (paper space linetype scale) to 1 and LTSCALE (linetype scale) to 1.0, or to 0.5 if you prefer more tightly spaced dashes.

I continue to prefer the modular one-detail- per-DWG file approach for the reasons that I gave in the July article. But I recognize that it’s not the best approach for everyone. My co-columnist Michael Dakan and I have encountered the multiple, overlapping, full detail grids method in at least two different design firms, so this new faction of the paper space cadets may be on the march. X-ruffians beware!