Any risks with updating templates?

[Richard Morrison]

I update my office template with new versions of AC, but each update is based on the prior version's (in some cases skipping a couple of versions) template. For all I know, there is data embedded in there going back to AC 6.5.

Each updated template seems to work fine, but I wonder if I am asking for problems by not recreating a template based on the AC default template for that version. Has anyone experienced problems by reusing prior versions' templates as the basis for a new template, or am I worrying needlessly?

[Richard Morrison]

I spent the last couple of days dumping almost everything in my monument to templates. The file size has gone from close to 100 MB to around 8 MB. Additionally, I have determined that I am going to live with most of the "out of the box" standards that come with AC: WE, building materials, line types, fills, pen sets, and surfaces. They are good enough in general, and I can always add one or two as needed. I only need about 5 "starter" composites. Upgrading to the next version of AC shouldn't take more than a hour. Import layers, import favorites, import keyboard shortcuts, adjust a few MVOs, and I'm mostly done.

For me, the important issue was in realizing that setting up in advance so many possible things I MIGHT use -- while comforting in a sense -- was taking far more mental energy and time in the name of "office standards" than it saved. When I do a roof plan now, for example, I have a choice of 4 possible layers, rather than over 30 in the previous template for that layer combination alone. Just trying to decide on the right layer and layer combination was causing stress. (The previous template had 62 layer combinations!) I guess having that level of granularity available is comforting to some people, but now it feels like "what was I thinking????" Saving a view previously involved choosing just the right layer combination. Now I just get "close enough," turn on/off whatever couple of additional layers I might need, and live with the layer combination name "custom." Not a big compromise.

I feel like I'm waking up from a bad dream.

[Anonymous]

Richard wrote:
now, for example, I have a choice of 4 possible layers, rather than over 30 in the previous template for that layer combination alone.

I agree. The first thing I always carry across to the person I explain BIM in general and ARCHICAD (or, God forbid — Revit) in particular is that one needs to understand what they are really doing and why.
Buttons you push, templates you use etc. do not matter much if you don't get the full picture.

In case of layers, for instance, we should base our decision on only four defining properties (pertaining to ARCHICAD):

• They can be turned on or off to hide/show content.
• They can be locked to prevent accidental editing (although, Holtinked Modules work better in preventing accidental editing, and, if you use Teamwork, they also provide much more secure access control)
• Allow for intersection control among structural elements
• Can be turned into wireframe in 3D to make complex 3D views easier to read

This is it... It is not hard to see that you really need very few layers that will do exactly what you want. No more, no less.

Just as a layer-related example, all structural elements in a functional ARCHICAD project can be located on 3 layers (I'll use Britain's Uniclass 2016 as example, but other layer-naming classifications are no different).

A-EF_20-М-Structural_Elements (Foundations, Frames, Beams, Columns)
A-EF_25-M-Wall_and_barrier_elements (Walls, infill walls, curtain walls, partitions)
A-EF_30-M-Roofs_floor_and_paving_elements (Roofs, Shells and Slabs)

This is it.
Yes, I think you can safely have load bearing walls and partitions on the same layer in ARCHICAD, because you can always turn off partitions by invoking Partial Structure Display (if your walls are set up correctly with BIM tags and categories).

If you ever run into an issue where you absolutely must have an additional layer — create it then. Don't do that in advance. You might not need it.

Richard wrote:
turn on/off whatever couple of additional layers I might need, and live with the layer combination name "custom." Not a big compromise.

Naturally, you will need to save it at some point if you want a stable Saved View, but, again, don't do that before you need it.


p.s. Trivia question: Autodesk's Revit promotional material mentioned that they moved on from layers to a new best thing — no layers, and ARCHICAD and other BIM packages are lagging behind in this regard. Not true, actually. Revit layers are still there... you just can't chose them anymore... 🙂

[Link]

I'm getting a little concerned about what I'm reading here, so think it responsible for anyone else reading this that I (re)join the discussion.

I hope that my 18 year's worth of building highly customized templates for ARCHICAD users all over the world, and my years as a BIM Manager for a large firm may be considered in context with my comments. I also assume that since we have the only ARCHICAD Templates at the price-points discussed, that you may in fact be referring to the BIM6x Template Kits.

These template kits were created primarily for our clients (we are ARCHICAD Resellers) in the US who are just starting out in ARCHICAD, because we see so many new users either not use the software at all, start off on their own and set up their templates incorrectly or not at all, or buy someone else's template which also leads them down the wrong path. All of this makes them inefficient and ruins their experience using the software, which leads to poorer resales and a bad name for ARCHICAD.

But these kits are also designed for people who already have ARCHICAD and perhaps even their own template. Either way, anyone is welcome to download our free BIM6x Starter Kit for evaluation purposes. The kit comes with a 171 page BIM Manual that explains all the concepts and may even teach the reader a thing or two about ARCHICAD best practices, plus an example project to reverse engineer along with work environment profiles and libraries, etc. Everything they need to show them what can be done in ARCHICAD and put them on the right path. Afterall only then can they really focus on their architecture.

If anyone likes the Starter Kit, they can upgrade to the Launch Kit or the Power Kit. No obligations, no pressure, but the offer is there. Our aim is not to make a bunch of money from these, but to help fill in some of the gaps for people on their journey from first discovering ARCHICAD to becoming highly successful users, so this is not a plug for our products.

Creating these templates however literally took thousands of hours, so to suggest a user can learn how to create and maintain a template in three hours is a gross underestimation of what it really requires to understand the intricacies of dealing with a complex list of attributes that are forever being evolved by both GSHQ and our local distributor, both of whom we happen to work very closely with. Not to mention exploiting all the features that GRAPHISOFT intended the user to take advantage of, and delivering it in a way that appeals to a wide audience of users in a manner that has the least impact on their learning curve or customization process.

Regarding layers, I don't understand the fascination of trying to cull the layer list down to a minimal few. I think they should total less than 100, but layers are important. You mentioned earlier that layers only serve four purposes:

1. They can be turned on or off to hide/show content.
2. They can be locked to prevent accidental editing (although, Holtinked Modules work better in preventing accidental editing, and, if you use Teamwork, they also provide much more secure access control)
3. Allow for intersection control among structural elements
4. Can be turned into wireframe in 3D to make complex 3D views easier to read

So what about filtering by extension, using layers for scheduling criteria and fields, including layers as Find & Select criteria, assigning layer combinations to views, using layers for mapping during export to other applications, using layers to reserve elements in teamwork, mark-ups, label association, the list goes on. Layers are important. Maybe for more than what each individual user may require but we're trying to show off all of ARCHICAD's amazing capabilities here.

So I would suggest coming up with something that works for you, but I would be wary of new users cutting their layer list down drastically. Manage them wisely, consider all their uses, hide locked layers if you need to, but most of all use favorites to control your layers and consider using views with specific layer combinations assigned to them. Not only will this be intuitive and easy for any newcomer to understand and use, but if you pre-link the views onto layouts, you won't really need to deal with them much, allowing you to focus again on architecture.

Some other statements which show a lack of understanding or at least a misrepresentation of the intended workflows:

Alternatively, one can buy some predefined template he/she did not set up and do not really understand, pay 100, 200 or even 500 dollars for them and the only result would be that you spend 5 hours deleting all 500 linked views in the most expensive template, 2 hours in the one for 200 and about an hour in the one for 100. What is the point?

So, templates, especially extended ones (some of which are even commercially marketed for upward $500) will take an enormous time to clean up, move every preset elevation, section, publisher set etc. Just imagine editing some 400+ items every time you start a new job. Do you have better thing so to?

I believe our template system is as flexible as an ARCHICAD template is able to be. It's just a fact that ARCHICAD templates can only be so flexible (hence the original topic of this thread), but we intentionally provide a bit too much so that it can be quickly stripped down, and we provide instructions on how to do that. A subtractive methodology is easily better than an additive one IMHO. If you're taking 5 hours to delete the views and then you're doing that every time you start a new project, then you're doing it wrong.

I'm all for customizing templates to suit your business needs. That is our complete intention and we preach it every day, but let's keep it real.

Cheers,
Link.

[Richard Morrison]

Link wrote:
A subtractive methodology is easily better than an additive one IMHO.

This is what I had always glibly assumed, too, but I have come to believe just the converse, at least for myself. When you start subtracting things, the process forces you to consider each and every item you are deleting. Do you need it right now? Might you ever need it later? There is always the lurking fear that if you don't need it now, you MIGHT, and so the temptation is just to leave it in. Then, not only is the file size significantly larger, so is the visual clutter. You end up with a very complex set of views, layers, etc. that you are always having to scroll through.

Whereas, if you have a good set of favourites (I'll use the UK spelling), when you need another detail window, say, creating a brand new one is not really any more complicated (and may be less so) than dragging a pre-existing symbol over, adjusting its borders and name, and adjusting its view folder, and then verifying which layout it is on.

Now, all that said, there is a certain minimum amount that MUST be set up in advance, but for myself, the glittering promise that you just set things up once for the largest possible project that you MIGHT be working on, and then the whole process is virtually going to automate itself, hasn't worked out in reality, at least for me.

With great respect for your contributions and knowledge, as well as being mindful of your business model, this is just my current thinking and what I've realized to be true for my own architectural practice. This practice is now residential only, and might involve anything from a tiny bath remodeling to a new house. In years past, I did hospital projects, and even go back to the time where there was only a single layer of graphite. Layers are useful, no doubt, but I'm not sure that we need the granularity of the AIA CAD layer system based in AutoCAD, which gave us:

A-FURN Furnishings or....
A-FURN-FILE Furnishings: file cabinets
A-FURN-FIXD Furnishings: fixed in place
A-FURN-FREE Furnishings: freestanding
A-FURN-PLNT Furnishings: plants
A-FURN-PNLS Furnishings: system panels
A-FURN-SEAT Furnishings: seating
A-FURN-STOR Furnishings: system storage components
A-FURN-WKSF Furnishings: system work surface components

When I look at the big list, I need to think about what I could delete. Especially for a beginner looking at this, there is a tendency to leave the whole list in -- just in case. But should a table go on "freestanding" or "work surface components"? On the other hand, if a template just has a layer called "Furnishings" and you come across a need for more granularity (and I'm not actually convinced you ever will), it's easier and faster to just add a single layer on an as-needed basis than to have to perpetually delete layers - thinking about the contingencies of an as-yet undesigned project each time. The mental bandwidth required to deal with the full number of layers makes me tired just looking at the list.

This is all to say that I think an additive approach can be just as effective as a subtractive one, especially for certain types of offices, and in some cases, more so.

[Da3dalus]

Consequences of template management become exponentially more important with the number of users you have in a firm. We have 80 employees, maybe 50 whom are regular ArchiCAD users, and half of those just know enough to be dangerous. Pre-populating Layers, View, and even standard Details is the only way to ensure consistent, or even legible output.

That being said, the last few version upgrades of ArchiCAD did not require a great deal of template updating. 18-19 involved almost nothing new, which is good, because I've been really busy this year! AC20 is pushing me to revisit the Favorites, since it is a vastly improved system, and will be used. We've been mulling over Favorite improvements for years, but now is the time to jump. However, Attributes, Layers, Views, and Layouts will probably not change this year. Time is finite; you have to pick your battles.

We are in the process offering more embedded standard Details. ArchiCAD's Detail process is dismal, and until I can drag-and-drop a typical detail (like our ubiquitous ADA-Compliant Fire Extinguisher Cabinet with Fire-Rated Wall option) from one project to another, maintaining scale, name/ID, and options all in one step, embedding is a time saver. Copying editable details is painful.

Being a large firm, we have to be somewhat Democratic. Though I am the Template Czar, I don't want to make major changes without some input. I am not infallible, and simplifying something that I don't see as useful could very well botch a workflow that out Healthcare or Interiors Teams depend on.

We also have a long history, with ArchiCAD projects extending back into the late 80's! Our Layer and Wall Composite systems have remained relatively consistent for a decade, which is a lifesaver when resurrecting old projects (it is common for us to get a job on a school or hospital that we last worked on 15 years ago). So, as much as I would like to tweak some of our systems, it necessarily has to be a careful, glacial process.

Finally, one more note about Layers. Quite a few reasons have been noted, and here's another: AutoCAD. Many of our consultants still use a 2D CAD workflow, and that is more than sufficient for small projects and remodels. The money we save may buy us a new computer, or furnish a room for our client. Layers are really the only was to universally identify drawing elements in CAD, and we can't spend a lot of time explaining to our engineers what line means what. By adopting the NCS/AIA CAD Layer Guidelines (layers like A-WALL-INTR), there's a universal language that ensures that, even when our models are dumbed down, much of the information remains intact.