NO MORE DRAWINGS

And when you need a / the3D model?

There's probably a huge (but mostly unrealized) market
for data migration services.

One major issue: knowing what & when to toss obsolete
or never-to-be-again-needed data.

And what of new releases that don't support some of the old
data features? Seems like a probem (sometimes regression)
for some systems.

What of systems that don't support data from two+ releases
back (they really only need to migrate the last some think ...)?

Or your vendor folded ... consider their customers & all that
data ... and systems that must have regular paid renewals
to even function ...

In 10 years do you think old format PDF files will be good
on your then-current hardware & software?

 

Joe,
My earlier glib answer - 'drawings' - has been niggling in the back of
my brain all day. Sorry about it, because you brought up a good
point. Drawings convey the CTF dims, but CAD commonly drives the
rest. So how do I save that to use in the next decade?

What if I try to open that CAD file in ten years and tool off of it?

With a SWx file, I won't warranty that the 3D data was the same if you
open it in ten DAYS if there was a new service pack to SWx. Most of
the baseline features will rebuild correctly, but there is plentiful
history of other features (lofts & sweeps for sure, and possibly fills
and boundaries based on their relatives) not rebuilding the same from
service pack to service pack.
And certainly, over the last ten years of explosive growth in the
product, we can expect 'fillet' and other features to act differently
after a rebuild today than they did in 1997.

If you want to retain your CAD data as a snapshot in time to represent
the released state of your design, I don't think (and in my experience
I know that) you cannot currently rely on history based files.

If you want to build the same tool in 2017 that you built in 2007 you
need a copy of the data in a 'dumb' format like IGES or, preferably
for SWx, parasolid.

And here I am just being Ed - To be honest, I have no data that those
files also won't change as versions change, but I make a point of
saving all released parts in both on the assumption that they are
locked in time. It's the only option that I know of that I have (love
to hear others comments on that)

_________________________________

To predict the next ten years, all I can call on is the last ten
years.

Parasolid and iges were there ten years ago, and I have heard nada
about them going away, so they seem like a safe bet (except for not
KNOWING that the file as opened today is exactly the same as it was in
1997. Anyone tested this?).

The bigger problem in my mind is storage media... For instance, I
have hundreds of files on zip disks and I have had to make a point of
carrying over my old zip drive to access them How much longer will
that old zip drive be supported? Burned CD's turn into coasters in
less than half a decade (though there are exceptions in both
directions). Hard drives are cheap and seem to hold up for a good
long time, but who hasn't had personal or secondary experience with a
hard-drive un-recoverably failing? In ten years, will we still be on
hard drives or will it all be solid state? No clue.

Then there is turnover in the company (which hopefully a pdm system
will mitigate, assuming that your current pdm is still viable in 2007
and that 'Bob' followed the rules) and a lot of other soft factors
(when 'Bob' left the company did he tell someone how stuff was
saved?).

I don't know what will happen in 2017, but maybe we can extrapolate
lessons for the future from what happens today.

When we are asked to reproduce a design from 1997, the CAD data is
usually lost on some zip disk or a long lost tape or hard-drive, and
everyone involved has moved on to other jobs.

1. Give me the drawing so I can understand the CTF dims and
tolerances
2. We 3D scan an early article of the product (or in the worst case
the worn tooling of the prodcut) so we can rebuild it from scratch.


In 2017, It'll always be a bonus round if we have an IGES or Parasolid
of the orignal nominal design, but in our back pockets we will always
have the fallback of 3D scan of early articles or tooling for lost
data files. But for CTF dims and tolerances, nothing in my
imagination will ever beat a drawing. And that is the long version of
my earlier glib answer.

Hope this has some value to you,
Ed

 

I'd not expect to be able to use a 10 year old ParaSolid file in
5 years without migrating it upwards release by release thru
most/many prior releases *which you must have in functioning
status on functional *for them* hardware* to do so.

Hence migrating the old data with every new release (or close to it).

Though some UG version 4 data once migrated to version 10+ <g>.
Had a few problems with "corrupt" surfaces from version 8 or 9
though ... user errors of some sort I suspect .... later versions
would not allow such.

 

What if I try to open that CAD file in ten years and tool off of it?

This got me thinking too. Long-term storage is a problem. My first
thought is to suggest an alternative to your suggestions for a long-term
format. Parasolid is still an evolving format, so I'm not sure it would
be a good idea to trust it. IGES is fixed by a standard, so it should be
as stable as long as the translation software builds models correctly.
I've had better luck with STEP files, but YMMV. STEP is also a
standardized format, and it is still being developed. Not so much
changed, as added onto. I'm not aware of any standardized format that can
store texture data though, and that might be important to some designs.

This problem is larger than just the CAD industry. Governments are also
taking an interest in long-term storage of documents, and several have
realized that MS Office can't be trusted. This is why there is so much
news in the IT industry about MS trying to force adoption of their
"open" format as an ISO standard. OpenOffice.org is a real threat to them
in this situation. IIRC there is already a non-proprietary ISO-standard
xml formatted document format, and OpenOffice.org 2.0 can use it as the
default format. In other words, future compatability is guaranteed.

Drawings have nothing that guarantees future compatability with an
official standard. The closest thing is pdf, which is the de facto
standard, but still proprietary. It's based on PostScript, which is at
least published. I'm not sure if it's an officially adopted standard. DXF
is also well published, but it shares AutoCAD's limitations, which are
numerous. (The most insurmountable reason dwg/dxf compatability can be a
problem is that there are things that the format cannot store, which
other CAD formats can). TIFF is, I believe, a standard, but obviously not
ideal. Still, it's as safe as the storage medium.

I agree with you that hard drives are the most reliable backup tool. I
almost said magnetic medium, but I never met a reliable tape backup. The
problem with storage of electronic data is that it is so fragile, much
more so than paper. Printing drawings and raw model data to microfilm
might be the most stable thing to do, but I've never heard of it being
done. I think the most secure thing to do right now would be keeping
miltiple copies on physically separated hard drives that are very
regularly refreshed to new hardware. Recent reports show that the most
common age for HD failure is two years (for running drives), so that
might be an important figure in how often a HW refresh is required. At
thre very least, the drives should be replaced with whatever the current
mainstram interface is. PATA is going to be completely gone soon,
replaced by SATA drives. I don't know whay you would do right now if you
had critical data on a SCSI RAID set. I don't know enough about that
interface to know what forms can still be used.

At any rate, any company with data they want to keep should obviously
have a data storage system in place. The one advantage of electronic
formats is that it's so easy to make a backup, unlike those rooms full of
flammable drawings they used through the last century. The trick is to
actually make sure that you make those backups.

 

The data migration issue & support for it is a prime
consideration for most CAD/CAM vendors.
But they also usually assume users have clues.
You can add new features but must still somehow
support the old.
Some folks have added their own new features & new
entity types now & then to some systems but these then are
usually not later vendor-supported, may not migrate well (if at all)
without the source AND Harry Programmer ... PLUS the rest of the
software may not work with them.

Consider adding a "triangle" entity as an example .... just to
annoy jb <VBG>.