hi,
i’m from the polish chapter and we start a workgroup for the creation of the national construction classification. in order to get a picture i have a few questions:
as i gather the rest api is the way to access bsdd from any local national construction classification level. how can the producers and suppliers acquire the bsdd guid for their products’ specifications? do they have to program the access to the ifd.* domain themselves?
probaby will get cleared on our internal meetings, but i’ll ask nevertheless: when we choose a future classification standard, will we have to implement the bsdd connection, or will it be made on the hq-level?
ok, @3 i’ve figured out myself, it’s not the same guid, as also every pset has their own guids, apparently useful in the future relations, according to the parellel discussion on another bsi forum.
So far, the only “shared” aspect of classification seems to be the adherence more or less to ISO 12006-2 which is better for some classification systems than others.
As for a shared “format” for classifications: there is none, but when Graphisoft introduced the ARCHICAD classification system, they came up with an XML format to store and exchange configured classifications. This is not IFC-based, but a few software applications have imported the Graphisoft-prepared classifications from (https://www.graphisoft.com/downloads/archicad/BIM_Data.html) into their own systems.
Might be a candidate for standardisation. However, be aware that there are no GUIDs used whatsoever, so it is purely based on strings expressing a code/number and a descriptive string.
as for the classfication standards actually there is one, outside of archicad: this is the classing system in vectorworks. in the uk the vectorworks’ class names bear the uniclass 2015 names, and the hierarchy is being supported using the ‘-’ character in the class name. you can have up to 4 subclasses in a folder-subfolder look, then the names are in one row.
To my knowledge, Uniclass at the moment is good, better than others, but still has some issues, they added some extra attributes like IFC entity names, etc, but it caused their classification be messy
I think IFC needs has a standard classification for itself as well, while supports multi-classifications
I’ve started a practical project, currently called: ODClass, to find some alternatives (which is not standard yet, and even I changed IFC structure to find a better way to support my existing projects)
For instance, IFC in Controls has “Scope, Cost, Time (and resource separately)” which is equivalent to Iron Triangle, but I want IFC supports Agile Triangle which IFC needs some expansions to support “Quality” as well
Also, I’m thinking to move some entities like Site, Building, BuildingStory, etc, into Context (or relevant name entity for context to support “where” in my new classification)
The bsdd guids for properties, propertysets, quantitysets do match those within the definition of IFC. Some of the published properties using an enumeration are inconsistent in published IFC versions although should be corrected for future releases. Can you please clarify how you came to the conclusion they are different?
hi guys,
thanks for your feedback, but i think we’re drifting a bit here. what i’m mostly interested in is the way the national classifications use bsdd for the ifc connection.
@jonm
what i’ve meant was the fact that those guids were different strings.
i’ve stumbled upon an ustrian site (freebim.at) where there was a hub for programming of the properties for the objects for the austrian bim norm a-6241-2. the available methods list retrieving of the bsdd guid ( string getbsddguid() ) and setting that guid ( void setbsddguid(string guid) ), so i’ve assumed they were independent from the global ifc guid and ifc properties’ guids.
The bsdd guids for properties, propertysets, quantitysets do match those within the definition of IFC.
In my understanding, the GUID in IFC, i.e. Pset_BeamCommon in IFC4.2 is “3vyMG0qSCHuO00025QrE$V”, and it matches with the GUID in bSDD as shown the search service of bSDD like:
it would completely reverse what i’ve observed. in this case the retrieving of the bsdd guid would be for the blank object’s ifc data, and the setting of the bsdd guid would populate bsdd data with the already created ifc guid for some object, so the other way 'round from what i assumed.
As Yoshi posted, you can find a bsdd propertyset Pset_BeamCommon
This has a globalId 3vyMG0qSCHuO00025QrE$V
We are about to flush and recreate the Ifc doc definition refactored for version control (but this link should persist). You can find the same UniqueId here
I have seen previous publication of property templates might have inconsistent ids, but publications from now on should have this resolved. I’d suggest to check the .ifc or .ifcxml included in the draft schema for IfcRoad extension for a reliable listing.
Both have same GUID, so anytime IFC GUIDs be correct, at the same time bSDD/IFD GUIDs would
Publishing resources/publications I don’t think be time-consuming
I think it’s better we find a reasonable solution for “unique article identities” too, like GTIN
GUID inherently, especially in Relational Databases, causes increasing file sizes in databases, so adding GTIN will double the issue
I think If bSDD GUIDs are a reference, then GTIN can be integrated with GUIDs to join BIM world to GS1 world and vice versa
bSDD or IFD as I know develops with Catenda AS under bSI support.
But IFD needs more than raw data and templates generated based on IFC schema
For this reason, bSI cooperates with GS1, but even GS1 has its pros and cons:
For instance, GS1 is good at some product templates and has GTIN and is active to develop some Internet of Things - IoT solutions mainly with their Korean community
But even GS1 has some cons, their product templates don’t cover “models”
bSDD/IFD should be a 3D model that data/information mounted on it and even has material/texture
thanks, ehsan.
i’m in contact with havard bell from catenda, but from what i gather it’s still not clear who is in charge of the tying of national construction products to the ifc data via bsdd (in order to use the bsdd guid for the uniqueness and for mapping to other classifications)…
and yes, a great idea of the bsdd multidimensional data model. will it require the particular pset guids as well?
but there must have been the first appearance of the guid, either in the ifc object, or in the bsdd prototype, or am i missing something? i can’t believe both guids were created simultaneously.
how about getbsddguid or setbsddguid then?
i presume i still don’t get the exact mapping procedure…
