Saturday, April 28, 2007

knowledge security

BIM post a big threat to authors.  Yes, it is true if the database itself is not properly managed.
But this seems true too for paper-means communication.  People panic with the security issue of IT side more just because they have little knowledge about how the data were manipulated inside the mahcines and through the wires.
Drop the hardcore hacker cases first.  Today many of the advance CAD-based software builders have developed a second data type for sharing purposes.  These light-weight format vs. propriety format help protect the geunie knowledge and methodologies about how the contents were built.  Usually they only contain essential data enough for the down-stream user ( e.g. contractor) to do their own work.  A very common example is DWF developed by Autodesk.
Some people refer such technology to pdf format of acrobat reader.  I would say they are far more "safe" in terms of intellectual property security.  pdf format are about graphic and text, which could be easily transformed to other formats by making screenshot and text recognition softwares (many of them bundled with scanners you buy), or simply decoded by ametuer hackers.
In graphical world, potable format files only contain approximate data accuracy (e.g. 0.1mm) - for example, 2 lines are seen joined in a potable format are actually broken lines.  Such low-feidility data technology reduce the file size and hinder the data reusability at the same.  Design-wise, rule-based or knowledge-based features were also thrown away - for example, the depth-span ratio of a steel beam is optimized for a designer-style staircase.  This render the digital data dispatched similar to those in legacy paper era - not better, not worser.
But this is only part of the story, since you have to make sure the data you shared and forming part of BIM is good for the purposes of the whole project team, there are many good chances that they conflict your security policy.
We'll discuss this issue later.

Tuesday, April 24, 2007

More on Details

People have knowledge about building and BIM would tell you to draw necessary details enough for general purposes of BIM.  Their reason: "value".  So, it's a waste of resources to hire a crew to draw the anchor bolts of a kitchen cabinet.  Also there're other reasons like software/hardware performance, data query and analysis performance, capacity of machines ....etc.

While it's too ture to argue on this issue, our mind become too solid to take a further step or just to put a fullstop at the current prospectus of BIM.

But just ask one question: are there eventually any drawings showing the hanging bolts supporting the kitchen cabinet at the end of the day?  Surely it's "yes" - not only the fabricator, but also the project engineer have to approve it's structural adequacy based on the drawing duly prepared!

There must be  "advantages" to put any data about a building into a BIM, where they appear and are referred to during a building life cycle for some time.

This is the axiom of BIM!

What we should have asked is "if we make effort to include them in BIM, is there a positive value generated?"

That's why we have a balancing point about what to/not to build in BIM.  So currently, we wouldn't expect to see an 3D anchor bolt in a BIM.

But the "balancing point" itself is not the axiom! 

That is, it's existence and position is not always true.  It should be moved or even vanished when BIM technology keep advancing so that we take lesser and lesser effort to put more and more details inside a centralized database.

The future of BIM is not hard to imagine, because it's fairy clear that we don't need the chemical equation of a waterproof membrane to appear in a building project.

Friday, April 20, 2007


I went through a few occasions briefing practitioners about BIM. There's a common "phenomena" I noticed from their understanding: - "Well, we've to push the design development to earlier stages with these tools", and "our practice should make change to accomodate this advance technology" ... This "understanding" is more deepened if they "tasted" the softwares by themselves.

As I've addressed before in "Objects, Domain Objects....": in principle there's no hassle to change our process to suit particular tool if you find no justification at all.

Subject to blame: Objects. Yes, objects again - the god and devil in one body.

People start "tasting" BIM by playing with the domain objects: walls, slabs, windows, doors, stairs ... mostly. Relevant or irrelevant they also found many details of the objects they didn't expect to think of during design stages (which they assumed when playing) e.g. finish thickness, stair nosing or railing styles.

Consequently they regard the details as a "must consideration", concluding that the "details" should be taken care of at earlier stages.

My suggestion: Master your process before master the softwares.

If it's your thinking that it's a waste of time to take care of certain detail at certain stages of design, you should look deeper into the software and ask: "How?"

Generally there are 2 approaches:

1. Modify the object's parametric template to delete the unneccessary details

2. Leave them alone as "dummies". These "dummies" could be a reminder to alert your colleagues to take care of them at later stages, or as a "backup" in case you run out of time to finish them up.

If you take the second approach I will assume that you have a organized strategy to document your design processes. You may also ask how to differentiate the "dummies" from "true details". I suggest to use the 4th syntax of a 3D model i.e. color, material, transparency, etc. Take a good look into the softwares and you should find these object properties.

Thursday, April 19, 2007

Draw, modify, coordinate, schedule

Whatever you will do modeling a building, especially build up a BIM, every elements should go through these 4 steps.

To draw is the easiest and toughest task. It's very easy to build anything in 3D you want. But to make it a reasonably easy object for the following 3 steps - modify, coordinate, schedule, is a kind of state-of-art.

To use the pre-packed libraries comes with the BIM softwares will help you through this. However, if that's a 3-dimensinoal hyperbolic skylight in your mind, then you must prepared to spare more time on this "draw" step. The best way is to go through all the 4 step as trial run to make sure your home-brew object is good enough.

But life is not like that. There are cases you have to change objects you made to an extent that you have to make a new one to replace it. Check the software and your drawing convention that you could do that revert/remake in a non-painstaking way.

As an example I give here the least possible trade of building elements in the world a designer would like to coordinate (or even to get a peek) - "builder's work" shown in a CBWD (combined builder's work drawing). Yes, a hole in a wall/slab for a pipe to go through! But they mean money to builder if they're not coordinated.

But I was told to schedule them out from a tower BIM into excel sheet!!

Like everybody I was scared with this idea. Instead of blaming the guy who ask the question so late (BIM was built and modified for almost a year), I turn my full attention to the software function to see if this could happen with minimal working hours.

Luckily I found that there's a feature of the software to put the results of a "measure" command into an object, like doors or windows. Eventually I can schedule out these "measure objects". What my draftmen have to do is:

- convert all "void" geometries (objects that making up the holes in walls) back to positive solids. It's a one step procedure for a whole floor. That means a 70 steps for a 70-storey high tower

- 2 "measure" commands for each discrete solid, so we could have the surface area and depth of each particular builder's work. It's painstaking but it's routine. 2 working days were spent for the first 15 floors. (including learning curve)

In the end, I spent half the time of traditional method (use calculator and CBWD hardcopy) to come up with the schedule, with the outcome could be reused for other revisions. Also, it's about 1/5 of time to reorganize the BIM to schedule out the works in normal way.

Conclusion : life is different from theory, but they all start with theory.

Wednesday, April 18, 2007

Automation, Collaboration, Consistency, Solver

BIM come with many advantages over our traditional methods of doing things - from inception, design, contracting, construction and facility management.

People come to you to describing the capabilities of BIM by explain the features of BIM - coherent database, parametric modeling, collaborative client-server strategies, bundled libraries, popular data formats, allied partner softwares/solutions, work scheduling tools, seamless diagnosis applications ... etc.

At first instance you may map the above BIM features/software features to your particular business processes.  Instead I would suggest you first look into the nature of these improvements if they are your/your company's concern, in terms of cost saving and quality assurance:

1. Automation
What are the value generated if particular human process was replaced/improved by putting it into machines?  Is there any balance you will strike for Quality vs. Time vs. Cost?  What and where is the learning curve?  Is there any new loopholes generated by eliminating the old loopholes?

2. Collaboration
Is quality collaboration crucial to your business?  Differentiate them from in-house collaboration to external collaboration and treat them separately.  You will probably get a 100% "yes" from your boss to both streams because it's a standard PR answer.  I assure you it's not quite so if you put this issue in monetary terms.

3. Consistency
Practically we never get 100 out of 100 marks offering our services.  One of the major reasons is work consistency - drawings, specification, illustrations and even site works.  Is "consistency" giving you a competitive edge of your service?  Give it a measure and balance it with your input to achieve it.

4. Solver
Works could be very complex that you have to put more people, and hence, money to work it out.  In extremity you may report to your boss that it's just not feasible at all to place tender on certain works.  Today's softwares open you new opportunities to tackle complex jobs and render them "feasible".  Have a good study of them to "level up" your company's profile in faster pace.

Once you categorize your improvements in these 4 fields of application, you could review it easily with the same language.  Avoid making matrix tables of evaluation for sake of global understanding.

Saturday, April 14, 2007

People, Workforce, Professionals

Say you just acquired a driving license, your best friend lent you a Porsche, with turbo, nitro, rocket propeller(??) ... whatever, and told you: "drive it work, dating, shopping, to breath country air. You'll forget how to live without it."

Yeah, you forget because you may probably lose you legs - that is, you have to forget.

Turn to any page lectures implementing IT solutions, there's always a ROI graph telling you the cost and return versus time during implementation. Not only investment, time and opportunities may suffer the same for design house or contracting companies, making the input potentially higher. Luckily the majority of AEC-related businesses are human-force-centric, there are always affordable plan B and team B to tie up the lose ends if anything fails.

Implementing BIM won't cost you limbs.

Many of the failing, or unsatisfactory stories I came across were related to people:

- people see innovation as career threat

- enthusiastic people have no time; those who have were forced to be enthusiastic

- fluent software operators are scare

- ambiguous target user group BIM softwares were designed for

- rapid staff turnover

While we have little control over some of these hurdles, there are some golden (at least I think) rules we could follow to make BIM a success, regardless of these glitches:

1. BIM is about efficiency and quality of works which involves a series of processes and methods, together with software solutions.

2. Before you make any decision to choose what software to use and define your goal, make sure you have setup guidelines or procedures to make sure the process and methods you gonna have will be recorded down and managed from time to time.

3. These guidelines is like a PQP of ISO-9001, except it's just a few lines of principles written on 1 piece of A4 paper.  Because whatever BIM software you adopt should have a particular instrument (software capabilities) to help you record and maintain these procedures.

4. If your belief is similar to mine that BIM would streamline the design documentation process and coordinations, the procedures will form a major chapter of your office's PQP, though they may not be legitimately so.  Be serious about it and put your seriousness into an action plan, otherwise people will just find hard to follow you. Worst of all, they think you're just waxing the company logo.

Just like you don't want your people to take away the T-bar when they leave, make BIM like all other stationaries to stay with your company.

Thursday, April 12, 2007

Objects, Domain Objects, Object Orientated

Todays intelligent software vendors will tell you you'll be using their softwares to "build" model, not to "draw" model. They also tell you they have a whole set libraries of "objects" accompanied with their solution for you to build a virtual building in a flash of light.

"Trash it or boss fire me!" illustrator thought.

"No, their library don't suit my practice." boss said.

"Are you sure they legally comply?" architect asked.

"There's no hyperbolic-shape wall? It's crap!" designer commented.

Me sarcastic? No! Not even a tiniest sense.

If a piece of software is not designed to work your way, you expect the whole company work their way around to suit the tool, would you?

All 4 people spelled out their concerns very right. Unfortunately at the same time they made certain preconception about the capability of current softwares designed for AEC industries.

Software providers used the term "Object" or "Domain Object" because it's a direct reminiscent of the underlaying programing philosophy: "Object-orientated", or "OO" they used to build these softwares.

In computer world anything can be "Objects": numbers, letters, words, cars, men, cats, values, series of electronic pulses, actions, properties, behaviors ... ... anything that can be abstracted.

When software developer invented "AEC domain objects", they did not merely create parametric model templates so that you could build very fast models. (please read my last chapter titled "Parametric Modeling")

Instead they create "generic object" in first hand: wall, slab, column, window, door, envelope, etc. These "generic objects" do not have geometric particularities yet. So "wall object" could eventually take the form of a horizontal waffle; "slab object" could have a stick-like form.

So what's the meaning and use of these object if they are so "generic"?

Answer: properties and behavior.

Some examples:

1. A window object could not exist alone, it should have a host object to adhere to e.g wall object

2. A slab object will be attached with "floor area" property, while wall object would not.

3. A load path should follow a column object, not envelope object.

With the definition of these very basic classes of object laid down in our AEC world, the software developer start to build up the semantic convention to describe everything (almost) about a building.

"Why tutoring me with these boring talks? I'm not interested in IT anyway!!"

I'm not taking you into IT world (actually I'm not capable), but only to let you understand more about the tools at our disposal.

So, we can see now, the "library of objects" people referred to at the start of this article is merely a second layer of pre-made instruments we could make use to speed up our modeling work.

We could actually build our own "second layer" library based on the "generic object class" provided by the softwares.

The example I showed in "Parametric Modeling" is a custom "second layer object" I built myself. I'm an architect, not programmer. I have been using CAD-based softwares for 12 years.

Here I provide a concept only. How far an application open up the interface for us to build our own domain object varies from software to software.

It's always the hard part to put things into practice. I will cover more practical cases in later blogs.

Wednesday, April 11, 2007

Parametric Modeling, Knowledge Capture, Knowledge-based Modeling

These terms are frequently referred to when talking about BIM technologies. But few people know the concept and power they enabling us. So I'd like to share my experiences started with this topic.

In a nutshell, they are all about intelligent modeling methods.

To build up a 3D model, BIM model, whatever you call them, in an intelligent way:

1. We should input as less as we could, leaving the repetitive input (drafting) works to the machines - computers.

2. The outcome should be easily modified and managed with lesser effort as possible

In this sense, you will no longer bother which software is CAD(drafting), CAD(design), BIM, 3D/4D/nD modeling - people accustomed to label the softwares they like or dislike with these terms.

To me, how good a software is to what extent it enable us to achieve point 1 and 2 as delineated above.

Back to the terms, Parametric Modeling is the technology itself, it can be used as a general term, or a feature of softwares. Knowledge capture, knowledge-based modeling is the intelligent method we could deploy when modeling, enabled by this technology.

The following is an example: a typical GFA exempted bay window 3D model popularly found in majority residential development in HK. All the instances were built using one single parametric template.

All you have to do is to input the:
1. Width of window
2. Which wall it was located
3. Setting out in plan view

The machine it will give you the cantilevered structures, window sub-divisions, window panes, window frames, wall opening, etc

Why did the computer know what and how to draw all the items you didn't specify?

Because it already know the bay window should be designed in such a way so:
1. To exploit the view (means money in HK)
2. Maximum cantilever span is 500 mm
3. Minimum sprendrel size at top and bottom of the wall allowed by Law of HK
4. Maximum glass pane width is 1400mm (structural concern in common-sense)
5. Enough openable window pane for cleaning purposes (600 mm wide each)
6. 1100mm high protective barrier should be provided for openable pane.

Why it knew?
Because some human being (with professional knowledge) input it beforehand, and these input were re-used when you make your own piece of bay window.

This is what we call knowledge-based modeling. Knowledges and rules are stored (captured) and reused in subsequent modeling procedures.

Is that GOD?
Depends. There may be twenty-something types of bay windows used in a residential tower, but there may only one type of toilet window in the same building. So don't waste time to capture knowledge about that toilet window if you don't find it's style useful to your other building designs.

P.S. This example was done with Revit 5/6, email me if you wanna give it a shot

Tuesday, April 10, 2007

Any Questions

I love questions. Especially questions about BIM.

You can post your questions here about BIM and could expect answers particularly on:

- BIM generally

- Change management

- Autodesk Architectural Desktop

- Revit

- Catia

- Digital Project

- Virtual Reality

- Building


- 3D/4D modeling, parametric modeling

- Detailing

- anything so that I could gather answers from the community, expanding this blog as a knowledge base

What's this range about?

A brief what-about could be referred to "About" at the right column of this blogger page.

Everybody who interested in BIM are welcomed to comment on this blog or mail me directly.

Thanks first your taking time to read, and hopefully, comment this blog.


Hello world!

Hello BIM Technology Range!

Hello me.