Have the original Mathsoft patents lapsed, and will there be any competitors soon?

[PhilipOakley]

The original Mathsoft/Mathcad patents for the live maths whiteboard were many years ago. Those patents prevented the likes of Maple and Mathematica producing similar 'live math white board' products, and they were all stuck with having to do the 'FORmula TRANslation' approach of the text expression with after the fact symbolic displays.

Looking at (US/WTO) papetnt lifes it looks like the lifetime of those patents was 20 years http://patents.stackexchange.com/questions/312/how-long-are-patents-valid, so anything before Jan 1996 should have lapsed. Mathcad - Wikipedia, the free encyclopedia shows that the original mathcad was released in '86, and by '96 Mathsoft was already up to V6.0 plus, so most of the key innovative aspects wre already there (some still swear by V6!).

There have been more recent changes such as the text entry change, but that isn't part of the basic 'white board'.

Is it likely that there will be other live maths whiteboard providers any time soon, especially extensible ones that can fix the gripes?

[radair]

I haven't looked at the patents for a long time, but one of them was storing a 32 bit pointer to data in the available bits of a double precision NaN.  This lets one efficiently store different data types in an array.  Interestingly, that still works with 64-bit pointers on Intel hardware because it actually only uses 48 bit pointers which can also be stored in 64 bit NaNs.  Someone could have screwed up the patent timeline by patenting that now that a 64 bit OS is the norm.

Robert

[PhilipOakley]

I've looked out some of my old manuals.

The V3.1 for Mac doesn't mention their patents.

I don't have my V6 / V8 manuals to hand.

but the V11user manual [(c) 2002] lists US patents:

5,469,538; 4 Mar 1994 - Mathematical document editor and method performing live symbolic calculations for use with an electronic book.. 

5,526,475; 2 Mar 1994 - A mathematical document editor that can perform live symbolic calculations. The mathematical document editor is capable of placing..

5,771,392; 23 Jun 1998 - A method and system for encoding nested matrices and vectors. The system can be used in a live mathematical ..
5,844,555; 1 Dec 1998 - A method and system for locking a defined portion of the live mathematical document. A live mathematical document ..

6,275,866; 3 Mar 1998 - Software components such as objects, applications and computational blocks are linked through a data flow manager and an ..

It looks like 5,771,392 (http://www.google.co.uk/patents/US5771392) is the one you refer to. That one is not 'out of date' yet relative to a 20 year rule.  (check with your local lawyer - I'm not in the US !)

Looks like I've mentioned this before in Units - who needs them!‌ (try to find 'patent' on the page/thread - may need the google cached version!)

So, so have lapsed, but others are still valid.

[RichardJ]

Patents filed on or after June 8 1995 last for 20 years from the filing date (not the publication date or the grant date). Patents filed before June 8 1995 last for 20 years from application date or 17 years from the grant date, whichever is longer. So the 1994 ones are both dead. The "encoding nested matrices" patent will expire on June 20th this year. I didn't bother to look up the other two.

[radair]

I should probably have added that even if Intel hardware used 64 bit pointers, the method would still work.  Half the memory is reserved for the OS which defines one bit in the data address space, and since the pointer points to a structure, it will be at least 8-byte aligned in memory which fixes 8 more bits.  That gives 9 bits for NaN encoding which is plenty to handle IEE NaNs like infinity and implement custom NaNs like "this is a pointer that needs to be left shifted 8 bits"  I haven't looked to see if there are any patents, PTC/Mathcad or otherwise on the topic.  I have no idea how Prime handles the 64-bit OS issue.

Robert

[radair]

ummm.... 8 bytes actually fixes 6 bits, but that's still ok.  I just did a quick search and ran across this which explains the scheme in detail  Pointer magic for efficient dynamic value representations making it unlikely that there are hidden patents.

[PhilipOakley]

Thanks, that's a good blog on how it's done (in various other systems). Looks like mathsoft-PTC aren't fighting for that patent right 😉

One of the other patents looks like 'MathConnex', and another looks to be the Area locking mechanism - I was one of the folks pushing for some of that functionality - particularly the ability to copy out of a locked sheet when that sheet is a corporate reference sheet.

Philip

[JasonC]

Given that the WISYWIG (live math whiteboard) patents have lapsed, why haven't Maple and MATLAB added it to their products?
Or was some kind of extension filed?