Data Protection

I don't think you can write protect the file as you describe, but a checksum should be all you need--you don't really care WHY the file changes once it leaves your hands, just IF it changes--and you can tell that from the checksum

Not once you lose control of the official single source part. The systems administrator can do about anything to almost any protection scheme.

Possibly you could use public key encryption on the part file though.

Just altering the system's date is easy.

Which would do you no good.

How long do you need protection? That's a *major* issue. 20 years?

Best bet is probably to write the file to three CDROMS (I think the media may last 20 years; it's iffy) and pray there will be equipment and computers that can read them in 10-20 years someplace. (I recall that the raw 1980 census data was unreadable after less than 20 -- the only computer in the world that had a slight chance of reading the tapes was in a museum in Japan without an OS or power IIRC. And it's very unlikely that magnetic tapes will last 10 years though things may have improved.)

Send one copy to your customer.

Anyway, send one copy in a sealed, notarized (be certain to have the notary mark the seal so is would show if alterd) envelopes to yourself AND an old-line law firm thet's likely to still be in business in 20 years no matter what. Send these by certified (or the most expensive mail with signed receipt back). Put your unopened copy in a safe deposit box (including both receipts). Have the law firm do the same. Pray . Better send two copies to everyone .. and hoe one will still be good in 20 years.

HTH

the multi-cd approach is probably your best bet. some things to help with keeping records that will last: read this website:

the cd's at a lower rate - don't burn at hyper speed. why? bob z. is just paranoid about this one.

another informative site about cd-r's

I agree with Michael's suggestion, having used it myself . I always hand off files in a zip archive. Winzip calculates a CRC checksum for each file in the archive, and by backing up a copy of the archive you have access to the checksums if the 'originality' of a file is questioned.

Art Woodbury

SNIP

Not really. Certainly, in some cases the sum will change but not in all by any means.

A binary compare to the original (*certified* original) is proof of change. And how are you going to store the checksum for 20 years?

Don't forget: IF you alter the part you have on disk it's been changed and is useless as proof.

Speaking of suchthings, how are you going to read & display the file in 20 years? Some seem to be having problems with parts from

2000 already .

How are you migrating your legacy data from two releases ago? Longer than that? It's probably pretty useless if you do not ...

you store the checksum by writing it down on a sheet of paper-- it's about six digits long. Storing that for 20 years shouldn't be any big deal.

Changing the binary file will change the checksum (there are some exceptions to this, but I think we can neglect them--and there are more robust methods if you really care).

And you don't need the original file as proof--all you need is proof of what the checksum was at the time you delivered it--a notarized letter from a witness should be sufficient. It goes back to my original point--it's not Chris's responsibility to find the changes if the file was altered--he simply needs to be able to demonstrate whether it was altered or not.

If the customer chooses to migrate the file to CAD systemX (and that includes SW2005) two years in the future, he does so at his risk--the original engineer is not liable for any translation errors that may result

I'll also concur with the idea -- but be sure you include the checksum in a message to your client(s). That way they can't come back to you later and blame you -- they can see for themselves that the checksum with their modified model doesn't jibe.

'Spork'

Art Woodbury wrote:

bob zee wrote: . . .

...and you're to blame. You give love a bad name.

Philippe Guglielmetti Wrote the following.

"If you look in apihelp.chm under "macro feature security" you'll see: swMacroFeatureSecurityCannotBeDeleted (this is my favourite...) swMacroFeatureSecurityNotEditable swMacroFeatureSecurityCannotBeSuppressed if you create a macro feature with those options and call Feature::SetUIState(swIsHiddenInFeatureMgr) you could invisibly run any code........"

Maybe you could use this to create a Macro Feature that would not let any changes to be made, unless your username was found or something to that effect.

H'mm interesting answers from you all. I'll give it some thought and come up with some sort of scheme before I have to do this again.

Thanks to all,

Chris

I (and my lawyer) disagree--in the event that a design is modified, it is the responsibility of the modifier to document the changes, and demonstrate that changes are not significant.

A set of (unspecified) modifications after the part leaves your hands is essentially a "get out of jail free" card...

and if they hope to prevail with that point, they have to demonstrate what their changes were....

I'll stick with my current lawyer, but thanks for the suggestion....

Chris,

In my estimation, are not the DRAWINGS supplied adequate documentation to define what you as the engineer have specified?

You are providing models as a convenience to a manufacturer (or end buyer), which as I understand it is a GREAT thing. In my experience, when a model is shared, you as the designer have a better chance of getting the part you designed as you get rid of all "re-creation errors".

If something is added after your design for manufacturing purposes, this will be done regardless of your supplying models or not.

Some Ideas:

Option 1 - Supply data but with the the disclaimer that the models are being supplied for reference puposes only.

Option 2 - Write out PDF's of your drawings to document (and freeze) what was designed. Make this the "build" standard.

Option 3 - Supply a dumb format.

Option 4 - If you are making hybrid prints (i.e. paper print + cad model for undimensioned features), then document this in the print by mentioning a specific "dumb" file -> "Undimensioned features are defined by the cad file 12345.igs"

Option 5 - Never underestimate the value of sharing your best data with your builders. It will help them better make your parts as you designed them.

For the legality, I have no idea what is best. Maybe don't supply anything, including part prints, as it reduces your risk to zero. (grin)

Regards,

SMA

You have just killed the largest single advantage of 3D CAD/CAM systems by thinking of 2D paper as the end product of a design process.

The 3D model can be used all over the place downstream and measured as well at any point to clarify an issue. Or to make the part or tooling or ..

Many of the large firms, such as GM, use the 3D model as the actual controlling model and any dimensions shown on any prints made from it had best be right or a designer is likely to lose their job. And millions in scrap made.

The drawing is for annotation of specific items not well captured by the 3D model. If you are to make a part from such you get the 3D model and perhaps a drawing .... notes & tolerances have to be somewhere, sometimes.

By sending only 2D drawings the entire 3D part may have to be remodeled by many downstream ... all sources of more error & waste.

Nor are 2D drawings any good at all in defining surface data or machining.

Lazy designers on some systems have been known to just edit the text in a dimension as well ..... you can guess the typical rsults.

What about just exporting and reimporting to and from Iges or something. The advantage is they will have all the data but none of the features can be changed without adding new features. Any new features in their Properties will have Created by ThisPerson. Then if changes are made it is easy to prove that they did not match the data you sent them, and who made those changes. Seems too simple of a solution but I think it should work. Forgive me if somebody has already mentioned this.

Corey

I forgot to mention that that will make a dumb solid. Some people know others may not.

Hi Cliff!

If I ran the world (I don’t as last glance) I would abolish paper prints or rely on them to annotate things like material specs / finishes and critical to function dimensions. But in the today’s industry, I have yet to see many strong uses of the model-only paradigm outside of prototyping or a fully in-company process (particularly when a part is not designed and made at the same company things get worse). I have see good implementations of the CAD/paper hybrid.

I’ve been on all sides of it and have worked with paper only, paper+cad and cad only. Each has its benefits and weaknesses. I have worked as the guy building parts (and tooling as well) and the guy designing parts, so I understand the value of the model.

But, unfortunately paper still holds a very high currency for those among us who have not yet been blessed with CAD (estimators, purchasing agents, managers, painters, platers, inspectors (usually), etc.). The truth here is that a fully dimensioned print still holds the highest legitimacy for many people, not all of them engineers and manufacturers.

Regards,

SMA

I just though I might add this in response:

It is great to see that my all-powerful thoughts can still topple the world (Smile). Thinking is not dangerous. Tangible things are the end product of the design process as I recall, engineering drawings (and models too) - a means to that end. Perhaps this might not be what I think at all.

Cliff – You’re ok in my book!