Top 5 likes

[PhilipLeitch]

Okay � I�ve listed my gripes, but to be fair I feel I should give a balanced view and list the top 5 thinks I like about Prime (this time in no particular order).

1: The ability to copy equations out in text form:

(@APPLY (@LABEL FUNCTION str2num) (@ARGS (@APPLY (@LABEL FUNCTION substr) (@ARGS (@SEP (@SEP Date 😎 2)))))

Great. That means that I have a greater ability to share equations without having to save the entire worksheet. Text is such a simple medium that it works with email, SMS� almost anything. Changes can be made to the text and it will paste back in.


2: The grid and snap-to-grid look and feel. Although I don�t always want to use this (so I would prefer it as an option), it sure does make alignment much easier than Mc14. I know you can hide grids, but I would like to move equations and text where I want them � and that means not having a snap to grid option at all.

3: Plots. Okay� they aren�t �there� yet functionally, but they look more presentable. So when the features end up being implemented, the plots will be a winner for me.

4: Design-Of-Experiment (DOE). Although many of the concepts are really quite basic, it takes some of the tedium out of the tasks. More importantly, after the experiment is contructed the DOE features allow feeding in of data and presentation of results. So a nice addition.

5. Tables. I like the tables as presented. They still need a little work (try and paste in a couple of thousand rows from Excel to see what I mean). But all in all I think they are a good addition.

Philip
___________________
Nobody can hear you scream in Euclidean space.

[IRstuff]

Not only can you copy equations, but you can modify or create your own, potentially.

Presumably, once you figure out the complete syntax, you could yacc your way into a translator for other programs.


TTFN,
Eden

[Fred_Kohlhepp]

On 8/26/2009 7:39:50 PM, pleitch wrote:
>Okay � I�ve listed my gripes,
>but to be fair I feel I should
>give a balanced view and list
>the top 5 thinks I like about
>Prime (this time in no
>particular order).

>4: Design-Of-Experiment
>(DOE). Although many of the
>concepts are really quite
>basic, it takes some of the
>tedium out of the tasks. More
>importantly, after the
>experiment is contructed the
>DOE features allow feeding in
>of data and presentation of
>results. So a nice addition.

Okay, help out a poor old soul; how are these functions "design of experiment?" I'm used to experiments being measurements of physical quantities, maybe that's my problem.



Fred Kohlhepp
fkohlhepp@sikorsky.com

[PhilipLeitch]

Yes. It's design of statistical analysis of normally biological processes.

In sciences you do something and measure it. This is no different, just doing more things.

When you do more than one thing there is a chance that the multiple things you do have an interaction.

There is also the problem if there is large variability in measurements how do you determining when a difference actually does occur?

For instance, I want to know if my fertiliser makes my plants grow faster. How would I test it?

I can't treat just one plant and measure it to see if it grew better. Instead I have to treat a lot of plants and take a lot of measurements to find out average and variance to determine if there is a difference.

If I grow one crop, then fertilise then grow another crop I'm not controlling for the different seasons when the two crops grew. One might have grown in summer and the other winter.

So I might divide the field in half and over the same season fertlise one side and not the other. But if I do that the side I fertlise might be slightly higher/lower and therefore get more/less rain. Or perhaps the soil is different on one side to the other.

So I might divide the field into multiple plots, randomly picking which ones to add fertliser to and which ones I don't.

THEN... I have the option of also mulching.

A properly ransomised trial over the field will tell me:
1. Whether there is any net benefit to fertilising.
2. Whether there is any net benefit to mulching.

But wait, there's more. What if mulching has no great growth benefits BUT assists fertiliser absorption by the plant?

We can test for that and find:

3. Whether mulching AND fertlising provides MORE growth than would be expected based on the known effect of each treatment on their own.





Philip
___________________
Nobody can hear you scream in Euclidean space.

[IRstuff]

I have't looked at that section, but one of the main benefits of DOE is the evaluation of fractional factorial experiments.

In the case of n parameters that could affect growth, a classic experiment would be to test every parameter against all the others. If the parameters are multivalued, then the full experiment usually requires more resources than reasonably available.

Fractional factorial experiments allow you to combine parameter changes so that effects are still statisticallly observable, but without the full set of possible experiments.

If MP1 includes that, then that's a good thing.

TTFN,
Eden

[Fred_Kohlhepp]

So this is specialized statistical analysis.

Fred Kohlhepp
fkohlhepp@sikorsky.com

[PhilipLeitch]

Yes - I've seen it described it as "Research Methods", "Experimental Design" and now "Design of Expermiments"

It applies well to bilogical data, especially things like sociology, psychology and other "softer..." sciences.

BUT - it also applies well to things like engine testing. An engine can run differently depending on temperature of the environment, fuel used, load, driving conditions, and so on. So again you have potential interactions. e.g. You need a vehicle that does very well with Fuel A for longhaul driving, or an Engine X that is the best all round urban engine for a variety of vehicles.

But how do you know that Fuel A becomes unusable in hot conditions, or that Engine X will not perform well in low temperature conditions?

You would need to repeat your readings a number of times, controlling what you can (distance, vehicle type and load, Fuel, etc. and "replicating" the test in a variety of locations (i.e alpine, sea level) and time (summer, winter, morning, night) for those thigns you can't control.


So all in all it is just a process for eliminating factors that you can't control. The environment isn't and can't be controlled, leading to a lot of noise. Repeating the experiment will reduce the noise. The DOE features helps minimise the number of repeats of the tests you do while also ensuring that the effect of the uncontrolled variables is minimised.


Philip
___________________
Nobody can hear you scream in Euclidean space.

[Fred_Kohlhepp]

Thaks for the explanations.

Sounds like useful stuff if used properly.


Fred Kohlhepp
fkohlhepp@sikorsky.com

[josephbradley26]

On 8/28/2009 6:56:55 AM, fkohlhepp wrote:
>Thaks for the explanations.
>
>Sounds like useful stuff if
>used properly.
>
>
>Fred Kohlhepp
>fkohlhepp@sikorsky.com

Fred
It can be extremely useful. As a Lean Six Sigma practitioner, it is a core part of Projects that I work on.

Joe

[LarryLederer-di]

The five things I have found that are better so far in Prime are





1) The ability to work with complex numbers in both rectangular and polar format. big plus for an electrical engineer.





2) Increasing and decreasing decimal places by clicking on an icon (similar to Excel) is an improvement over going into a menu.





3)The combination for the text block and text box allows for better handling of text.





4) All the new built-in functions and constants are great. However, PTC will need to do a better job of letting the user know they exist and how to use them. If it wasn't for my old manual from Mathcad 11, I'd never know half of what functions are available.





5)The grid and the snap feature is something I would have liked to seen earlier. Still no simple drawing tool to go with it.

[TomGutman]

The snap to grid is not new. Mathcad has always aligned regions on a rather coarse grid. What has changed is the coarseness of the grid (from approximately half a character to approximately two characters) and the presense of visible grid lines.
__________________
� � � � Tom Gutman

[LarryLederer-di]

On 8/28/2009 7:38:07 PM, Tom_Gutman wrote:
>The snap to grid is not new.
>Mathcad has always aligned
>regions on a rather coarse
>grid. What has changed is the
>coarseness of the grid (from
>approximately half a character
>to approximately two
>characters) and the presense
>of visible grid lines.
>__________________
>� � � � Tom Gutman


Correct, but you had to guess on where you were putting things and over lapping was a major problem in getting documents to layout in a presentable way. This is an improvement over previous versions. By the way Tom can you find five things you feel they should keep?

[PhilipLeitch]

Tell me if I'm wrong, but wasn't the alignment after the fact? You couldn't set alignment on as you create new regions. If that's the case then the grid is a BIG change. Like everything, fixing formatting after the event is never as easy as getting it right in the first place.

But you touch on a really good point, the spaces are much too wide. Having the ability to alter the grid size would mean that you could make the grid very small and then hide it, which would have the same effect as not having a grid at all. I'd like that feature to be added. But if it were, we woud still need the "alignment" buttons.

This is all very much like the old Visual Studio form designer. It had to weigh up the pros and cons of having a grid. In the end most developers did use a grid, but made the grid increments very small, and STILL allowed users to align controls and adjust the exact location irrespective of the grid. This was over a decade ago, but many of those features are still in the current Visual Studio IDE.

Philip
___________________
Nobody can hear you scream in Euclidean space.

[TomGutman]

Alignment to the grid was always forced. In one sense, all screen based systems have a grid, the finest grid possible being the pixel. Mathcad's grid was about half a character -- a bit fine as grids go, but there wasn't (and, as far as I can tell, still isn't) any provision for positioning finer than the grid. Quite different from Word, where you can choose the grid spacing and also use the alt key to allow dragging to pixel boundaries. Word also allows positions to be specified by menu entries, providing for positioning to less than a pixel. Such fine differences can't be shown on the screen, but Word is printer, not screen, oriented and printers typically have resolutions of five to ten times the screen resolution.
__________________
� � � � Tom Gutman

[ValeryOchkov]

On 8/28/2009 7:38:07 PM, Tom_Gutman wrote:
>The snap to grid is not new.
Yes 😉
See
http://en.smath.info/forum/default.aspx?g=posts&t=100
Val
http://twt.mpei.ac.ru/ochkov/v_ochkov.htm

[ValeryOchkov]

On 8/28/2009 5:55:42 PM, Derbigdog wrote:
>The five things I have found
>that are better so far in
>Prime are

1) The ability to
>work with complex numbers in
>both rectangular and polar
>format. big plus for an
>electrical engineer.

It will be good to have posibility to switch off complex number for more speed calc, less memory need and less errors - in 90-95 % engineering worksheets we do not use complex number. Ecxel for example has not complex number!

2)
>Increasing and decreasing
>decimal places by clicking on
>an icon (similar to Excel) is
>an improvement over going into
>a menu.

It will be better to have not "decimal places" but common number digits in answer.

4) All the new built-in
>functions and constants are
>great. However, PTC will need
>to do a better job of letting
>the user know they exist and
>when to use them. If it wasn't
>for my old manual from Mathcad
>11, I'd never know half of
>what functions are
>available.

It will be better to have inexpensive base Mathcad and extra purchasing special function puck.

Val
http://twt.mpei.ac.ru/ochkov/v_ochkov.htm

[LarryLederer-di]

On 8/29/2009 4:22:06 AM, VFO wrote:
>It will be good to have posibility to
>switch off complex number for more speed
>calc, less memory need and less errors -
>in 90-95 % engineering worksheets we do
>not use complex number. Ecxel for
>example has not complex number!

Over half of my calculations require complex numbers. It is one of the main reasons I use Mathcad.

>It will be better to have not "decimal
>places" but common number digits in
>answer.

That is already built in.

[ValeryOchkov]

On 8/29/2009 12:21:09 PM, Derbigdog wrote:
>On 8/29/2009 4:22:06 AM, VFO wrote:
>>It will be good to have posibility to
>>switch off complex number for more speed
>>calc, less memory need and less errors -
>>in 90-95 % engineering worksheets we do
>>not use complex number. Ecxel for
>>example has not complex number!
>
>Over half of my calculations require
>complex numbers. It is one of the main
>reasons I use Mathcad.
OK! You can switch on complex number in over half of your calculations.

>>It will be better to have not "decimal
>>places" but common number digits in
>>answer.
>
>That is already built in.
How, where?

Val
http://twt.mpei.ac.ru/ochkov/v_ochkov.htm

[LarryLederer-di]

On 8/29/2009 2:29:29 PM, VFO wrote:

>>>>It will be better to have not "decimal

>>>places" but common number digits in

>>>answer.

>>

>>That is already built in.

>How, where?



My mistake, I thought you were talking about fixed decimal places when you were talking about what I would call significant digits. Yes having a set number of digits so that the results do not show a precision that is not really there would be nice.

[SteenGroðe]

On 8/26/2009 7:39:50 PM, pleitch wrote:

>2: The grid and snap-to-grid
>look and feel. Although I
>don�t always want to use this
>(so I would prefer it as an
>option), it sure does make
>alignment much easier than
>Mc14. I know you can hide
>grids, but I would like to
>move equations and text where
>I want them � and that means
>not having a snap to grid
>option at all.

They just have to learn from most graphics application about nudging (setting different grids), guidelines and snap/no snap.

Steen Gro�e

[ValeryOchkov]

1: The ability to copy equations out in text form:

(@APPLY (@LABEL FUNCTION str2num) (@ARGS (@APPLY (@LABEL FUNCTION substr) (@ARGS (@SEP (@SEP Date 😎 2)))))

>Great.
No! HaftGreat!
We can (could) do it in Mathcad 11 by using normal (not "polish") notation:
http://twtmas.mpei.ac.ru/mas/Worksheets/rkfixad_Stiffr_Stiffb.mcd

Val
http://twt.mpei.ac.ru/ochkov/v_ochkov.htm