Have recently seen a few designs done using "Free" tools that include complexity like 32bit risc CPU's + FPGA's and matched length DDR traces -- Would have thought this would have only been feasibile with higher end tools a few years ago. (These are even open source free tools)
As a potential buyer of PCB Tool software my three criteria are: 1. How easily can the user accomplish things swiftly with this piece of software? i e ease of use 2. What can it do in terms of design complexity? 3. What does it cost?
None of your graphs explicitly cover point 3.
Buy having "corporate complexity" in your diagrams, you partly answer one aspect of point 1., collaboration in a team. However it doesn't cover ease of use for the individual user nor for a team.
I find your second diagram easiest to intuitively read because the axes are labeled and easy to understand. Diagram three is interesting, but requires a lot of text to explain; for example, what is meant by "high end" "low end" "basic" and "mainstream?"
I would suggest using several diagrams, and not try to say it all with only one.
Comment #1: By "Design Complexity", do you mean the complexity of the design you are trying to process using the tool, or do you mean the complexity of the tool's user interface design? My experience has been that the more expensive the tool (e.g., you have to be an "enterprise" to afford it), the more complex the tool's user interface and the harder it is to use the tool. JMO/YMMV Also, each generation of the tool becomes harder to use as more features clutter up the menus and functionality, until it becomes pretty much unusable. According to Alan Perlis, this happens with all software: "Every program eventually becomes rococo, and then rubble." But it seems to happen faster with CAD tools.
Comment #2: When I started looking at the graphs, my rather quirky mind immediately thought of the scene in Dead Poets Society (1989) in which Robin Williams draws a graph on the blackboard to show a systematic way of rating poetry. He has a mischievous look on his face while a student reads out loud the introduction to their poetry text:
If the poem's score for Perfection is plotted along the horizontal of a graph, and its Importance is plotted on the vertical, then calculating the total area of the poem yields the measure of its Greatness.
Williams steps back to show a graph labeled P.I.G. He then tells them this attempt to analyze poetry is utter nonsense (or a word to that effect) and tells them to tear the introduction out of the book, which they then do.
To me, a fine piece of software is like a great work of poetry or music. OTOH, a lot of CAD software is purely pedestrian, so perhaps 2-D analysis may be of some value.
@elisabethsimon: "we use an enterprise level Mentor suite with a separate PCB design group that does the board layout"
Splitting the PCB design flow in this way is a common practice. In fact, I know some companies that don't do PCB layout, but only schematic capture. Then, they send the schematic to speciallized third-party teams that provide PCB layout services.
About the graphs, I prefer the two first ones, but I'm not able to choose one. I feel that the actual "curves" will arise once we place some tools between the axes!
I like the view as a continuium (second graph) although the first would work if you put a third category in the middle.
For category names:
Individual level (one person working alone on simple designs)
Small business level (small group working on medium complexity designs)
Enterprise level (large company doing complex designs)
Of interest to me would be PCB tools priced for individual or small business with the capibilities to do more complex designs.
In my day job, we use an enterprise level Mentor suite with a separate PCB design group that does the board layout. On evenings and weekends I use lower cost tools to do simple to medium complexity designs. Currently, I'm using DipTrace which seems to be a good tool for the price although I'd be interested in a review of DesignSpark PCB.
Blog Doing Math in FPGAs Tom Burke 23 comments For a recent project, I explored doing "real" (that is, non-integer) math on a Spartan 3 FPGA. FPGAs, by their nature, do integer math. That is, there's no floating-point ...