For a moment, ignore design quality itself. Look only at the workflow from design and drawing output through material purchasing, production, and assembly. What consumes the most time? The answer depends on how the process and tools are designed. Across different tools and systems, the essential question is: which actions are redundant?
Scenario 1: A low-frequency operation creates training and waiting costs
In a 100-person R&D organization, design must request a part number from the PDM server to obtain a unique key. Deploying the OA-based number request takes a week, with two hours of training every day. Users understand the operation, but requesting a number is not a daily task. They learn it once and then begin a two-week design assignment. Two weeks later, a missing part requires another number—and they have forgotten the procedure. The request waits for an assistant.
Scenario 2: Information silos and “copying it once”
PLM describes an information-silo problem. From the moment a drawing creates it, a part has a name and remains the same artifact throughout its product lifecycle. Yet nearly every system—warehousing, drawings, PDM—works independently. Design calls the part Flange 1. Warehousing does not know that name and copies it again, which is error-prone. This copy operation is one of the largest sources of wasted time.
If data can be imported, or
the source is one unique file,
the problem disappears completely.
A unique data source: from the part file to the server database
A colleague once asked what the weldment CutList in a part drawing is for. It fits this scenario. If the weldment CutList is created from a drawing, it depends on the drawing file. Drawing dimensions still come from the part. If the part is the unique data source, the unique-source problem is solved. Otherwise, managing one part can require handling two, three, or even four files.
Linked values in a SolidWorks drawing solve the source for properties. Dimensions also come from the part. The CutList source is still the part.
With cloud servers, all of these can come directly from one unique database table.
Consider this: some drawings write Quantity into the part.
Is the part then the source of Quantity?
If another identical part is added in an assembly, should Quantity be copied back into the part?
Can that part be used in other assemblies? Should Quantity change there too?
In SolidWorks, a drawing is only a view of the part—it is for viewing, not the source of truth. The essential solution is to eliminate the copy operation and the checking and reconciliation that follow. In SolidWorks, the data source is the part file. With a server, the data source can also be a unique database or data table.
Scenario 3: distributing drawings and BOM data
Consider drawing distribution for one project with 350 unique parts, 1,400 total parts including 5,000 fasteners, and 150 drawings. How many parts must be purchased, machined, or kept in stock? How many are standard fasteners? How can the exact data and drawings be distributed to the right people?
Machining needs flat-pattern drawings; design works with part drawings.
Can all information update automatically when the part changes? An isolated Excel file cannot do this and would introduce another copy step. An exported Excel file can do it, if export rules define exactly what is included. Exporting everything solves nothing.
Design thinking is most active while modeling. I know exactly what I am drawing, whether it needs 45# steel, and whether it belongs at a welding shop or a machining shop. This is the best moment to establish the data source. Cataloging 150 drawings afterward is when the thinking has gone flat. In my view, solving the problem and locking it down while context is fresh is the most reliable approach because there is only one point of focus; reconstructing context later is unreliable because there are 150.
The SWTDO approach: property tags and one-to-one file relationships
SWTDO uses property tags and one-to-one file relationships.
Drawing distribution
Apply property tags to parts in batches. In WeChat, tags select a group of contacts; in our plugin, tags select a group of parts. Parts correspond one-to-one with drawings and can also correspond one-to-one with PDF and DXF. Multi-body parts are the point to consider here.
BOM distribution
Batch-filter parts by tag, then build a BOM.
Now consider tagging itself. If every file must be opened and every tag entered individually—and there are more than a dozen tags such as Material, Source, Machining Method, Purchase, and Welding—the omission risk is high. Can I see at a glance which part has no tag?
Batch tags and Data Center
Our plugin supports selective batch tagging. A tag can be a fixed string, a function expression, a combination of elements, or information extracted from the part.
The large BOM table is both a read tool and a batch write tool. Data Center is also a source for small BOMs. File property tags form the unique data source that runs through the entire product lifecycle: design, drawing output, retrieval, BOM creation, and distribution.