State of Play: 2017/wk09

As mentioned in previous posts my Windows XP laptop computer had problems, first faulty power adapter, which I replaced, and then the backlight for the screen. When I purchased the replacement power adapter I also bought a low cost replacement computer just in case the power adapter didn’t fix the problem. The replacement computer however had Windows 7 on it, and has since been updated to Windows 10. I also have a small netbook.

The failure of the Windows XP computer, also represented loss of Office 2003, AutoCAD LT 2000, and Multiframe 14. The loss of office 2003 being a consequence of the covert introduction of OEM licenses by hardware suppliers, and thus locking software to a specific piece  of hardware. The loss of AutoCAD LT 2000 because it won’t install, and the loss of Multiframe because the software for the hardware dongle is not compatible with Windows 10.

The use of Office 2003 has been partly resolved by unwillingly opting for Office 365 subscription. I consider the new software to be relatively unstable and to have a more cumbersome and inconsistent interface: the loss of the multiple document interface (MDI) is  a major defect. Unfortunately LibreOffice does not have a BASIC programming interface with the same capability has Visual Basic for Applications (VBA). So stuck with MS Office for the time being.

As for AutoCAD and Multiframe, I could run from my netbook, but the screen is too small and it lacks memory. Fortunately my Windows XP laptop, supports an external monitor, and I have been able to continue to run the software, however this is inconvenient, as most other software is on the Windows 10 computer. So it’s only a short term solution.

I do not believe it is necessary to keep software up to date. I hold this belief because the upgraded software is seldom ever a true update: it’s a different product with different capabilities. To explain by analogy: a screwdriver is not an upgrade of a hammer. Screws may have better holding capability than nails, but they also serve different functions as do the tools used to install these fasteners. A nail gun is not an upgrade of a hammer, it is an optimized version of a hammer, specifically for installing nails, it has none of the other functions of a hammer. similarly a yankee screwdriver, or spiral ratchet screwdriver is not an upgrade of a screwdriver, it is once again an optimised version of a screwdriver and lacks the full utility of an ordinary screwdriver. Likewise a power driver is a further optimised version of a screwdriver, lacking the full utility of a traditional screwdriver. The introduction of these optimized versions of these products should not result in the traditional tools being completely removed from the market.

But software is different! No it isn’t! Microsoft has changed the user interface of its products, and so the user no longer interacts with the software the way they previously did. New hierarchies of command collections need to be learnt, such problems do not exist with a proper command language and a command line interface (CLI). Now one of the benefits of AutoCAD over other CAD applications is the command line and its command language: but still Autodesk changed some of the functionality of the graphics editor user interface (UI) and thus user experience (UX). For example where previously one click would end a command, it was extended to click/search/click. Whilst the extra command options may be beneficial in some situations in the main they represent bloat and a decrease in productivity. Fortunately this functionality whilst set as default could be switched off to backtrack to previous mode of operation. Since LT 2000, the advertising for the new products, whilst introducing some features which may be beneficial, never really sold me on the idea of upgrading, not the least of which it cost too much.

As for Multiframe, I largely paid annual maintenance fee on the basis that it was the only Australian product that was heading any where near the capabilities of Master Series Master Portal and Bentley RAM Build, and unlike these products based on Australian standards. When Bentley took over, I wasn’t impressed with the way they contacted me, drowning me in documents and some rubbish about credit points, proposal to double my costs, and otherwise requiring my hardware dongle back to move to internet based licensing. I can figure where that was going, the optional annual maintenance fee will ultimately evolve into a mandatory annual subscription fee: don’t pay the subscription then don’t get to use the software. {We also have MicroStran license, but it seems that was also taken over by Bentley. Actually at this point in time, I don’t know who owns or sells the software: which I hazard was Bentleys intention, get source for Australian standards and merge with their own product and abandon the others. Which basically leaves SpaceGass as the only Australian structural analysis software}

For the most part I avoid using Multiframe, it is a bottleneck to the whole design process, and I prefer to do all my calculations in a spreadsheet (originally as-easy-as, then Quattro Pro, and now MS Excel). When we started we used our own in-house plane frame analysis program, and this is being developed further. Unfortunately it is not 3D and does not do non-linear analysis.

Though I am not certain of the value of nonlinear analysis,  I just wasted around 2 weeks trying to get a model to solve in Multiframe. All I did was extend the length of an existing model, and failed to converge. Besides a few other minor changes to the model, I had to change the solution iteration from a typical setting of 100 to a setting of 10,000. Fundamentally nonlinear analysis for plane frames typically solve with few problems, but once a series of such frames are assembled into a 3D building, and bracing is included for longitudinal wind loading, the solver starts to have problems. As far as I can tell the tension only bracing is too slender and flexible, along with associated struts, and this has a carry on effect to the rest of the structure. This tends to push up the required stiffness of the braced end bays, which the linear analysis does not require. Unfortunately AS4100 steel structures code requires nonlinear analysis (in Multiframe at least) whilst AS4600 cold-formed steel structures code does not.

So have hit a software problem which was not supposed to experience. The basic philosophy I held was that we should produce our own software for those tasks we are typically meant to be capable of solving with pencil and paper. Unfortunately where software was once optional and merely speeded things up, codes of practice have been modified to such an extent that techniques which require a computer are mandatory. For example residential footings are typically designed using slog or cord: as far as I am aware the vast majority of people using such software, whilst they may have a rough idea of the basis of the software, they would be unable to replicate the mathematical modelling of the analysis engine. Thus a dependency on software by others is becoming increasingly necessary, unless we move to spending considerably greater period of time in education learning the theoretical basis of such software and writing own analysis engines.

However formulating mathematical models and writing an analysis engine is only part of the exercise of developing software. Once reason for moving away from writing programs in Turbo Pascal and Turbo C, was the increased use of MS Windows, and the relative ease with which applications could be created with Quattro Pro for Windows, and the still greater capabilities of MS Excel combined with VBA. By using a spreadsheet, we can spend more time on the mathematics, and less time on the data file formats, creating and editing data, and reporting. Though a lot of time is still spent on presentation of reports in terms of getting information onto a page. For example changed computers and for some reason now need to reformat all spreadsheets to fit on the page because they are now too wide. Which seems to me there is a flaw in the relationship between screen presentation and printer, and document definition in MS Excel: the definition of the document should be based on required dimensions relative to the paper size used, and these dimensions converted for screen display. The width of columns in Excel however is based on some strange formula, relative to the default font. So sharing spreadsheets in the same office, with different computers is not so simple.

Anycase it does seem that the future is best served by open source software. Unfortunately at this point in time a lot of the open source software is cumbersome to use, which consequently takes time away from the primary task at hand. So until such software is convenient to use, stuck using the available proprietary products

I don’t see any value in upgrading AutoCAD LT 2000. It’s value will diminish as parametric modelling, and building information models are far more efficient once such models have been created. Such software however is currently far too expensive for limited use. With expensive software basically have to expend time pushing work through the software, rather than providing the original service got such software to assist.

Put simply with Revit and Tekla structure, should be able to define simple sheds, and canopies once, then for each project adjust a few parameters to get the new custom model of the building. So whilst it may take considerably longer to setup the parametric model, once the model is defined the time taken to complete a specific project should be considerably reduced compared to traditional methods: on condition that the new project does not introduce new parameters. Most of the structural assessment, design and certificates we issue are because shed and carport manufacturers have introduced additional parameters and thus sold something beyond the scope of their standard designs.

Unfortunately, the problem is that parametric software is expensive, and a user may only have a single parametric model. The next problem is that custom designed software to create a parametric model will likely cost more than a single Revit or Tekla license, and adding extra parameters with the passage of time will add further to the costs. Or will it?

If use Revit or Tekla someone still has to build the parametric model, and that takes time. Drawing a line and setting constraints is likely faster and easier than my calculating a series of points in Excel, and making one point dependent on another, then plotting those points on an XY-chart. Then further wrapping the points into AutoCAD script commands and running in Acad to create an editable drawing file, and thus also creating a drawing that can be correctly plotted to scale (unlike the Excel XY-chart). But then again maybe some add-in can bring a more intuitive parametric sketching and modelling capability to MS Excel. Furthermore when it comes to CNC machine tools and 3D printers, to what extent do we need a visual model? If we know what the object is meant to look like, and all we want to do is modify the parameters, then all we need to is generate the required code for the machine tools and printers: we do not have to display what it looks like.

We already have our 2D plane frame analysis program working in MS Excel and plotting moment diagrams and such on the 2D geometry. Making use of MS Excels shape layer to generate drawings from a few parameters is not that complicated, the main issue is the somewhat distorted scaling in the vertical and horizontal directions. However, if there are applications where presentation, and the scale on paper is not that critical, then it is a starting point, and if it becomes common place, then it will likely lead to the availability of more CAD oriented drawing objects. Alternatively there is LibreCAD, FreeCAD and OpenSCAD. Maybe LibreOffice Calc can be used to drive one of these CAD packages. I already drive Acad LT (command line scripts), IntelliCAD/ProgeCAD (COM), DesignCAD(COM) and Multiframe(COM) from MS Excel.

If I don’t need to use the graphic editor then I can still automate ACAD LT on my netbook, whilst I don’t have MS Excel installed there, I do have visual studio 2005 installed. Though the netbook screen too small for running visual studio and a lack of memory also makes it cumbersome. However can use WSH and VBScript. ProgeCAD on main laptop is ok for scripting though requires some changes to scripts compared to ACAD LT. ProgeCAD has a cumbersome user interface compared to Acad LT 2000, its main defect is lack of screen refresh after executing commands thus poor feedback that a command has executed. So I don’t like editing in ProgeCAD: also I am not sure if there is a lack of response in my keyboard or ProgeCAD object snaps are defective: they seem to jump all over the place or just plain exit.

So the gist is that I push MS Excel to its limits, and otherwise get others to do the work that requires new drawings or 3D nonlinear frame analysis: alternatively find online software services which can be used on an as needs basis.  As noted above I rarely need Multiframe, though I have greater need for Acad, on the other hand I haven’t had to draw anything new for 2 years. So CAD may be not all that important either.


Revisions:

  1. [03/03/2017] : Original