Teststand lite

[PhillipBrooks]

---

Santiago D wrote:

We have had a few requests for a less feature-rich version of TestStand. What specifically were you looking to add, remove, or modify? What would be the most important features to have?

I would also like to open up this thread to anyone else who would be interested in commenting on what they would like to see in a TestStand Lite.

Santiago D

National Instruments

---

I believe that the basic concept of what TestStand Lite would include is outlined in my post in the LAVA forum and subsequent entries made by other (much more prominent members) of the LabVIEW community at large.

What might be nice? A version of TestStand that would only sequence LabVIEW functions, and that would be part of the Professional edition of LabVIEW. The run-time for this version should be free. Maybe this could spur TestStand sales for NI? LabVIEW programmers could use this and learn it's potential and then justify the cost of buying the full package.

My posting on LAVA was made in 2006.  NI has made some added functionality to allow for licensing of toolkits within LabVIEW (Internet Toolkit, Report Gen Toolkit, State Diagram, etc...)

Maybe the time is right to consider a TestStand Lite Toolkit development license that would be part of Dev Suite Base and include only LabVIEW and  DLL adapters.

[Spectre_Dave]

Phillip

That is what I what. 

TestStand is wonderfully capable and complex.  It is like writing a memo with MS Word when all I need is notepad!  The only reason I used TestStand was because of a very short deadline that was imposed on the project.  We still had to contract out the TestStand work to a Certified TestStand Developer.

[MichaelAivaliotis]

You can already do this in LabVIEW. It's called a State Machine.

[Joe_H]

If all you are using TS for is to sequence LV functions, then just use a simple State Machine in LV. Then you have LV calling LV. Teststand is very usefull when you must interface different programming languages together like we do here where I work.

[Spectre_Dave]

I can't see myself using TestStand for any future projects. I'll write my own LV Sequencer and save a lot of $$$$$$$$$$$$$$$$$$$$$$$$$$$$$.

[Underflow]

Okay, zombie thread, but I'll toss in a use case.  (The following may sound grumpy.  Sorry ;))

I have a 3 axis robot running through an executable (standard state machine).  The user runs the robot around using the UI, and clicks a button to record a position to a "script".  Later, the user can play back that script.

I've implemented systems like this three times now.

It normally takes about 3 minutes after delivery for the user to request looping.  The script playback now needs to have a memory space for at least counters.  It also needs to implement conditional checking to know when the loop is done.

Next request happens at about 15 minutes: If/elseif/else statements.  More conditional branching. Pulling variables from the LV testing memory structure into the scripting memory structure (warning for those of you doing this: think about it.  Don't just branch that wire.).

Requests in the following weeks: Support for real variables, not just counters.  Support for ints, floats, and string types.  Full flow control.  Looks like I'm going to need a tokenizer, parser, lexer, syntax tree... wait a second.

Why am I building a language, again?  Aren't there enough of them in the world? And building it in LabVIEW, where there are no support tools for such silliness? 

Oh, yeah.  It's because there's no NI supported language where I can rationally call a single VI with a single line of text.  And every time I try to implement things in TestStand, I feel like I'm spending all my time trying to get around all the prebuilt  manufacturing cruft, drag'n'drop abstractions, and process model complexities.

LuaVIEW is *so* close to perfect.  If it were more stable, I'd go for it.  As it is, I've crashed the demo so many times, I can't currently justify it in a medical environment.

Here's my request to NI: Please tell me I'm wrong.  Please tell me I'm stupid and am missing something, and there's an easy solution.  I'm currently looking at the XML RPC server on LAVA written by mesmith, or possibly the LV Web Services and wondering if I can slip an open port past IT...

Joe Z.

[MichaelAivaliotis]

Have you tried Python calling LabVIEW VIs?

[Underflow]

---

Michael Aivaliotis wrote:
Have you tried Python calling LabVIEW VIs?

---

I've seen a system to do so, and considered it via TCP on the local system, but haven't done it yet.  Is that the mechanism you would suggest for parameter passing?

Message Edited by Underflow on 04-09-2009 08:54 AM

[mrbean]

I'm just now looking into using LabVIEW's Unit Test Framework.  I noticed your comment about using TestStand to call these unit tests.  Since I haven't used the UTF, can you elaborate on how TestStand would run these tests?  Is it similar to running labview tests?  This seems like a great way to exercise your code - a repeatable test procedure.

[Will_S.]

What you would need to do is develop the tests in LabVIEW and build a VI that runs them using the Unit Test Framework VIs. You should then be able to run this VI in TestStand.