does labview have a future?

[JimPanse]

Hello experts,

I'd like to make a general assessment. I have been working with Labview for about 20 years and occasionally contact Ni to solve technical problems or to find out about and buy products. I have been observing the trend that service has been getting worse and worse over the years for years now. I have not seen any real new developments in labview over the last 10 years. with the termination of nxg and the associated waste of resources, I have the impression that ni has to save money at every corner to keep the shop running. of course this is at the expense of customer support. i wonder if ni is about to go bankrupt or what is the strategy behind it? slowly i am looking around for alternatives to labview as i can hardly see a future for labview after the current impression. how do you see it...? what is your impression?

greetings, jim

 

[thols]

NXG was terminated, with a few corporate BS sentences, and NI then said they would add all the great stuff to LabVIEW in 2021(+), but nothing of that is in LV2021. I can guess that it is due to a time of turmoil until the resources can be aligned to focus on LabVIEW. But, if NI does not very clearly present the roadmap at the NI Connect event, I will be more concerned. However, I do not agree that no real developments have been made to LV in 10 years. Heck, 10 years ago we didn't even have conditional tunnels. I think the biggest new thing is interfaces, which is a crucial part of the language that has been missing. But then we also have things like channel wires, malleable VIs, and much more. But many other things about LabVIEW feel old, and that's where NXG was supposed to come in... About NI services I don't know since it has been years since I bought hardware or had anything else to do with them.

I cannot see any competitor to the productivity that is possible with LabVIEW compared to any other language. And the visual representation of parallel programming. So I hope NI makes the right decisions for LabVIEW.

[JimPanse]

Every now and then a small improvement comes along. a real innovation was the leap from version 7x to 8x. I think fpga is not bad either, but it is actually too expensive and there is no real support for it either. I work a lot with realtime and vision and I have to say that nothing has happened since usb3 vision. where, for example, are the coax press frame grabbers?  also that pharlab was discontinued without an alternative for desktop pc's. that's why i switched to labview linux and built the realtime system myself under linux. and if you ever need support for labview linux you're really on your own. in my eyes they've been milking the cow for the last 10 years... at some point even the last drop is sucked out. 10 years ago you called ni and wanted to test a card and 2 days later it was on the table. today you can't even find someone at ni who knows about stuff. NI sBRIO or System on Module was also a good approach.
but where is the interface to integrate smartphones into labview. or the pi... there are only non-commercial projects like linx.

[hooovahh]

1 hour ago, JimPanse said:

or the pi... there are only non-commercial projects like linx.

LINX now is supported on commercial applications starting in 2020 BTW.

Your opinion is valid, and you have reasons for it, but I think it might be a bit of forest from the trees situation here.  LabVIEW tends to have a one or two major bullet points of new features with each release, with many smaller improvements that are less noteworthy.  Some of these aren't very applicable to me and I don't see the benefit of the update, but I can still recognize that a bunch of effort was put into a making it into the release, and makes me think NI isn't sitting idle.  I know I made a list like this in the past when a similar topic has come up but I'm having a hard time finding it.

• 2012 - Loop Tunnel improvements with concatenating, conditional, indexing, and last value / Project Templates
• 2013 - Improved Web Services / WebDav / SMTP / Bookmark Manager
• 2014 - Actor Framework (I might be off by a version or two) / 64 bit Mac and Linux support
• 2015 - Custom Right Click Framework
• 2016 - Channel Wires
• 2017 - VIMs / Forward Compatible Runtime 
• 2018 - Command Line Interface / Python integration / Type Specialized Structure for Improved VIMs
• 2019 - Sets and Maps
• 2020 - Interfaces for classes / Free Community Edition with Application Builder

And here are a few of my favorite features that I can't remember what version they were added.  Error Ring, Improved VI calling under Application Control for starting asynchronous processes and static VI references, DVRs, Conditional Disables based on environment or Project variables, Linux Real-time operating system, allowing for 3rd party and open source tools to be installed and called with the System Exec, and then adding an embedded HMI, User Events, LINX toolkit for running LabVIEW VIs natively on a Raspberry Pi, or controlling an Arduino connected to the host, QuickDrop's plugin system allowing for all kinds of tools, filtering on search results, improved performance of tree and listbox controls, NIPM, and loads or more scripting functions with more added with each version.

I sure hope LabVIEW has a future because I've hitched my career to it.  But even if NI closed its doors tomorrow I feel like I'd still be using it for another 10 years or so, or until it didn't run on any supported version of Windows.  But I feel your concern, and if I were a junior engineer starting out, I would feel safer in another language.

[ShaunR]

48 minutes ago, hooovahh said:

LINX now is supported on commercial applications starting in 2020 BTW.

Your opinion is valid, and you have reasons for it, but I think it might be a bit of forest from the trees situation here.  LabVIEW tends to have a one or two major bullet points of new features with each release, with many smaller improvements that are less noteworthy.  Some of these aren't very applicable to me and I don't see the benefit of the update, but I can still recognize that a bunch of effort was put into a making it into the release, and makes me think NI isn't sitting idle.  I know I made a list like this in the past when a similar topic has come up but I'm having a hard time finding it.

• 2012 - Loop Tunnel improvements with concatenating, conditional, indexing, and last value / Project Templates
• 2013 - Improved Web Services / WebDav / SMTP / Bookmark Manager
• 2014 - Actor Framework (I might be off by a version or two) / 64 bit Mac and Linux support
• 2015 - Custom Right Click Framework
• 2016 - Channel Wires
• 2017 - VIMs / Forward Compatible Runtime 
• 2018 - Command Line Interface / Python integration / Type Specialized Structure for Improved VIMs
• 2019 - Sets and Maps
• 2020 - Interfaces for classes / Free Community Edition with Application Builder

And here are a few of my favorite features that I can't remember what version they were added.  Error Ring, Improved VI calling under Application Control for starting asynchronous processes and static VI references, DVRs, Conditional Disables based on environment or Project variables, Linux Real-time operating system, allowing for 3rd party and open source tools to be installed and called with the System Exec, and then adding an embedded HMI, User Events, LINX toolkit for running LabVIEW VIs natively on a Raspberry Pi, or controlling an Arduino connected to the host, QuickDrop's plugin system allowing for all kinds of tools, filtering on search results, improved performance of tree and listbox controls, NIPM, and loads or more scripting functions with more added with each version.

I sure hope LabVIEW has a future because I've hitched my career to it.  But even if NI closed its doors tomorrow I feel like I'd still be using it for another 10 years or so, or until it didn't run on any supported version of Windows.  But I feel your concern, and if I were a junior engineer starting out, I would feel safer in another language.

Nice list of features, not that I use any of them. Well. Maybe the occasional VIM but that was available unofficially in 2009 like many of the others  

You did, however, leave out the TLS for TCP in 2020 which is a huge one IMO.

Unless NI find a way to give us proper multilingual support; that alone will mean it's eventual downfall. I predict it is only a matter of time until LabVIEW becomes a SaaS because of that.

[Rolf Kalbermatter]

3 hours ago, thols said:

NXG was terminated, with a few corporate BS sentences, and NI then said they would add all the great stuff to LabVIEW in 2021(+), but nothing of that is in LV2021. I can guess that it is due to a time of turmoil until the resources can be aligned to focus on LabVIEW. But, if NI does not very clearly present the roadmap at the NI Connect event, I will be more concerned. However, I do not agree that no real developments have been made to LV in 10 years. Heck, 10 years ago we didn't even have conditional tunnels. I think the biggest new thing is interfaces, which is a crucial part of the language that has been missing. But then we also have things like channel wires, malleable VIs, and much more. But many other things about LabVIEW feel old, and that's where NXG was supposed to come in... About NI services I don't know since it has been years since I bought hardware or had anything else to do with them.

I cannot see any competitor to the productivity that is possible with LabVIEW compared to any other language. And the visual representation of parallel programming. So I hope NI makes the right decisions for LabVIEW.

NI didn't say they would be porting NXG features to 2021, but to future versions of LabVIEW. Technically such a promise would have been unfulfillable, since at the time the NXG demise was announced, LabVIEW 2021 was basically in a state where anything that was to be included in 2021 had to be more or less fully finished and tested. A release of a product like LabVIEW is not like your typical LabVIEW project where you might make last minute changes to the program while testing your application at the customer side.

For a software package like LabVIEW, there is a complete code freeze except for breaking bug fixes, then there is a testing, packaging and testing again cycle for the Beta Release, which typically takes a month or two alone, then the Beta phase of about 3 to 4 months and finally the release. So about 6 months before the projected release date, anything that is not considered ready for prime time is simply not included in the product, or sometimes hidden behind an undocumented ini file setting.

Considering that, the expectation to see any significant NXG features in LabVIEW 2021 was simply blue eyed and irrational.

I agree with you that LabVIEW is a unique programming environment that has some features that are simply unmatched by anything else. And there are areas where its age is clearly showing such as lack of proper Unicode support, and related to that the lack of support for long path names. Personally I feel like I could tackle the lower level part of full Unicode support in LabVIEW including full Unicode path support quite easily if I was part of the development team, but have to admit that the higher level integration into front panels and various interfaces is a very daunting task that I have no idea how I would solve it. Still, reworking the lower level string and path management in LabVIEW to fully support Unicode would be a first and fundamental step to allow the other task of making this available to the UI in a later stage. This low level manager can exist in LabVIEW even if the UI and higher level parts don't yet make use of it. The opposite is not possible. That is just one of many things that need some serious investment to make the whole LabVIEW platform again viable for further development into the future.

This example also shows that some of the work needed to port NXG features back to LabVIEW require first some significant effort that will not immediately be visible in a new LabVIEW version. While such a change as described above is definitely possible to do within a few months, the whole task of making whole LabVIEW fully Unicode capable without breaking fundamental backwards compatibility, is definitely something that will take more than one LabVIEW version to eventually fully materialize.  There are a few lower hanging fruits that can help prepare for that and should have been done years ago already but were discarded as "being already fixed in NXG" but the full functionality just for full Unicode support in LabVIEW is going to be a herculean task to pull off, without going the path of NXG to reinvent LabVIEW from scratch (which eventually proved to be an unreachable feat).

My personal feelings about the future of LabVIEW are mixed. Not so much because LabVIEW couldn't have a future but because of the path NI as a company is heading. They have been changing over the last few years considerably, from an engineering driven to a management driven company. While in the past, engineers had some real say in what NI was going to do, nowadays it's mostly managers who see Excel charts, sale numbers and the stock market exchange as the main decision making thing for NI. Anything else has to be subordinated to the bigger picture of a guaranteed minimum yearly growth percentage and stock price. The traditional Test & Measurement market NI has served for much of its existence is not able to support those growth numbers anymore. So they are making heavy inroads into different markets and seem to consider the traditional T&M market by now just as a legacy rather than a significant contributor to their bottom line.

Edited June 9, 2021 by Rolf Kalbermatter

[hooovahh]

1 hour ago, ShaunR said:

Nice list of features, not that I use any of them.

Yes, I don't assume you use many new features of LabVIEW from the last 10 years if you still develop in LabVIEW 2009.

[ShaunR]

20 minutes ago, hooovahh said:

Yes, I don't assume you use many new features of LabVIEW from the last 10 years if you still develop in LabVIEW 2009.

I almost changed with 2013 because of the JSON primitive but alas, it was as much use a chocolate fireguard.

I would change to 2020 because of the TLS support if it were not that I was forced to create my own solution 6 years ago 

If 2021 has multilingual support I will change to that but I won't hold my breath.

I'll let you whizz-kids find all the bugs in the latest versions and my crusty VI's will still work for you. That is the magic of LabVIEW. 

[hooovahh]

26 minutes ago, ShaunR said:

If 2021 has multilingual support I will change to that but I won't hold my breath.

You don't need to hold your breath for the answer.  There is a public beta, it can be downloaded right now for free, only requiring an NI.com account.  It probably won't surprise you to know the answer to your question.

[JimPanse]

I also think that it is the strategic staff that is leading labview in the wrong direction. the concept of labview is certainly unique and has its justification. the strategic decisions are simply the wrong ones. the management should be fired and a few innovative engineers should be left to make the decisions. there are enough new technologies waiting to be implemented with labview.

[Rolf Kalbermatter]

39 minutes ago, JimPanse said:

I also think that it is the strategic staff that is leading labview in the wrong direction. the concept of labview is certainly unique and has its justification. the strategic decisions are simply the wrong ones. the management should be fired and a few innovative engineers should be left to make the decisions. there are enough new technologies waiting to be implemented with labview.

That's always debatable. From a technical point of view I fully agree with you. LabVIEW is a very interesting tool that could do many more things if it had been managed differently (and also a few less than it does nowadays). For instance I very much doubt it would have ever gotten to the technical level it is nowadays if a non-profit organization had been behind LabVIEW. The community for LabVIEW is simply to diverse. The few highly technical skilled people in LabVIEW world with a very strong software engineering background, who could drive development of such a project in an Open Source model, do not reach critical mass to sustain its continued improvement. On the other end of the scale you have a huge group who want to use LabVIEW because there is "no programming involved", to parodize some NI marketing speak a bit.

Maybe just maybe, an organization like CERN could have stepped in just as what happened with KiCAD. KiCAD lingered for a long time as a geeky Open Source project with great people working on it in the typical chaotic Open Source way. Only when an organization like CERN put its might behind it, did the project slowly move into a direction where it could actually start to compete on features and stability with other packages like Eagle PCB. It also brought in some focus. CERN is (or at least has been) quite a big user of LabVIEW so it could have happened. CAD development moved in the meantime too, and while KiCAD nowadays beats every CAD package that was out there 20 years ago hands down, the current commercial CAD platforms offer a level of integration and highly specialized engineering tools, that require a lot of manual work when tried in KiCAD. Still, you can design very complex PCBs in KiCAD nowadays that would have been simply impossible to do in any CAD package 20 years ago, no matter how much money you could have thrown at it back then.

But LabVIEW almost certainly would not cross compile to FPGA nowadays, and there would be no cRIO hardware and similar things to which it almost seamlessly compiles to, if it had not been for NI. On the other hand, LabVIEW might actually be a common teaching course at schools, much like Python is nowadays on the ubiquitous Arduino hardware, if NI had decided that they want to embrace LabVIEW being a truly open platform.

The reality is, that we do live in a capitalistic system, and that the yearly earnings is one of the highest valued indicators for success or failure of every product and company. Could LabVIEW have been and being managed differently? Of course! Could it have survived and sustained a steady and successful development that way? Maybe!

 

Edited June 10, 2021 by Rolf Kalbermatter

[JimPanse]

I found the strategy of approx. 10-15 years better for my needs. if you continue to operate the current strategy like this, then I won't give labview much longer. that is the view from my field of application.... that might be enough for me until retirement.
i would be interested to know how high the profit per employee of ni was 10-15 years ago and how it is now. then you could estimate whether it is a successful course or whether it is going down.  

 

Edited June 10, 2021 by JimPanse

[Rolf Kalbermatter]

I don't have these numbers. What I know is that a few years ago, NI noticed that their sales figures were starting to flatten. For a company used to have high two digit growth numbers year after year this is of course a very alarming signal. 😀 They hired some consultants who came to the conclusion that the traditional T&M market NI was operating in simply didn't have left much more air in it to continue to support the growth NI had been getting used to. And strategic decisions were made behind the scene. Not to much of that has been openly communicated yet, but the effects have been quite obvious in the past few years. NI has deprioritized the PC based test and measurement hardware, completely abandoned any motion ambitions, marginalized their vision ambitions and put much of the traditional DAQ hardware into legacy mode. And their whole sales organization has been completely revamped. No field sales offices anymore, highly centralized technical support by typical call center style semi-outsourced places. Behind the scene they do large scale business and have increased their sales further since that alarming consultancy report. So somehow it seems to work.

One reason they may not have been very public about these changes is probably that it did change their old model of relying very heavily on external Alliance Members for actual application support of all their customers. In a few strategic industries they now have moved in to deliver full turn key systems themselves directly to the customer. For the typical Alliance Member that probably doesn't directly mean loss of business, since the customers and projects NI serves in this way are accounts that only very few Alliance Members would dare to even consider to look at, as the volume of the business transaction is simply very huge. However it certainly has other effects for all Alliance Members. The contact with NI has been getting very indirect with all the regional sales offices having vanished and the efforts from NI to compensate that with other means haven't gotten much further than marketing presentations with lots of nice talk up to this point.

As to LabVIEW: It's demise has been promised since it was first presented. First because it was clearly just a toy that no engineer ever could take seriously, then because NI didn't push for an international standard to formalize the LabVIEW G language, later because they didn't want to open source it. I'm not sure any of these things would have made a significant difference in either the positive or negative direction. It's clear that LabVIEW is nowadays a small division inside NI that may or may not find enough funding by the powers to be, to maintain a steady development. If you look at the software track record of NI it doesn't look to well. They bought many companies and products such as HIQ, Lookout with Georgetown Systems, DasyLab, and quite a few more, and none of them really exists nowaday. Of course a lot of them such as DasyLab were in fact simply buying out competition and it was clear from the start that this product has not a very bright future in the NI stall. Lookout was marketed for quite some time, a lot of its technology integrated into LabVIEW (LabVIEW DSC was directly build on top of much of Lookouts low level technology and the basis for the low level protocols used in Shared Variables and similar things). LabWindows/CVI is lingering a semi-stasis existence for several years already. It's development can't quite keep pace with the main contenders in the market, GCC and Visual Studio.

In view of this, acquisitions like Digilent and MCC may look a bit surprising. On the other hand it might be a path to something like a HP/Agilent/Keysight diversification. NI itself moves into the big turn key semiconductor testing business (and EV testing market), one of these other companies takes over the PC based LabVIEW, DAQ, and instrument control business.

Edited June 10, 2021 by Rolf Kalbermatter

[JimPanse]

so, if you look at the share price.

https://www.nasdaq.com/de/market-activity/stocks/nati

you can see a nice separation there. until 2016, ni was reasonably solid and had a slow steady growth. in 2016 they decided to make everything more efficient and lived at the expense of their customers. at some point the customer notices the reduced performance and looks for an alternative. which is reflected in a share price decline since 2018..... a bold summary... this roughly reflects the correlation of the share price with my personal experience with ni.

 

[Rolf Kalbermatter]

55 minutes ago, JimPanse said:

so, if you look at the share price.

https://www.nasdaq.com/de/market-activity/stocks/nati

you can see a nice separation there. until 2016, ni was reasonably solid and had a slow steady growth. in 2016 they decided to make everything more efficient and lived at the expense of their customers. at some point the customer notices the reduced performance and looks for an alternative. which is reflected in a share price decline since 2018..... a bold summary... this roughly reflects the correlation of the share price with my personal experience with ni.

It very much depends how much back you dare to look in that graph! 😀

Since it is a NASDAQ share you should rather go to the source, Luke! That steep climb around 2017 is actually after they started implementing those changes. The decline you see is mostly happening during Covid but as a trend not quite very significant yet.

That all said, the current trend to measure everything in share price is a hype that is going to bring us the next big crash in not to much time. My guess is that once most people have crawled out of their covid imposed isolation in their private hole, they will look at the financial markets and wonder where the actual real world value is, that some of the hyped companies would need to have, to make their share price expectations even remotely true. And then the big awakening happens when the first people start to yell "but he isn't wearing any clothes".

Edited June 10, 2021 by Rolf Kalbermatter

[Bryan]

20 hours ago, hooovahh said:

I sure hope LabVIEW has a future because I've hitched my career to it.  But even if NI closed its doors tomorrow I feel like I'd still be using it for another 10 years or so, or until it didn't run on any supported version of Windows.  But I feel your concern, and if I were a junior engineer starting out, I would feel safer in another language.

I'm in the exact same boat.  While I'm not on the cutting edge and doing much of the cool stuff many of you in this forum do, I do entertain the idea in the back of my mind that LabVIEW may one day go away while I'm still in the working world.

I'd like to think (realistic or not) that if NI ever decided to completely get rid of LabVIEW, that they would show appreciation to all of we who have made it successful in the past by making it open source.  

[hooovahh]

1 hour ago, Bryan said:

I'd like to think (realistic or not) that if NI ever decided to completely get rid of LabVIEW, that they would show appreciation to all of we who have made it successful in the past by making it open source.  

I hope for that future too, however I get the feeling that with 35+ years of LabVIEW development, that parts of the system are in a state that open sourcing the project might take a larger effort then NI would want to put into it.  Then there is the liability issue if some kind of exploit was found because the source was released.  I'm lucky enough to work in an environment where my boss asks for work to get done, and is less interested in the means of getting it done.  If I were making test systems for external customers I'd be more nervous, but internally I can just continue to use a legacy platform.  Cobol anyone?

[jcarmody]

After years of mergers, acquisitions and reorganization I've finally ended up in a Test Engineering group that primarily uses C#.  The current environment of legacy LabVIEW systems will keep me gainfully employed (maybe) but I'm glad to be facing the specter/opportunity of learning a new language.  Wish it was python, though...

[X___]

I don't know about the future, but seeing how dynamic the first public beta is, the present is not particularly bright 🙂

[JimPanse]

Is there an official statement from NI how long Labview will be supported?

[Neil Pate]

I think LabVIEW is going to be around in some form for at least another ten years, primarily due to how entrenched it currently is in the niche it fills. However, if I was starting my career again in 2021 I would definitely not choose LabVIEW as my primary language of specialization. I think the poor decisions NI have made over the last ten years will lead to its irrelevance/obsolescence, which I am genuinely saddened by.

It pains me to think where current gen LabVIEW could be if the money and time funneled into NXG had been put to overhauling current gen.

 

[Rolf Kalbermatter]

3 hours ago, JimPanse said:

Is there an official statement from NI how long Labview will be supported?

No, and there likely never will be! NI has a very strong tendency to never ever talk about what might or might not be, unless the actual release of something is pretty much announced officially. In the old days with regional field sales offices you could sometimes get some more details during a private business lunch with the field sales office manager (usually with the request to please not share it with anyone else, since it wasn't yet official and in fact might never really materialize in that exact way).

The only thing we know is that the LabVIEW source code was in some escrow and supposedly still likely is, for the case NI might end up with its belly up. How much that really means, other than to sooth managers who were worried about choosing for a single vendor software solution, I can't really say. It certainly doesn't mean that the LabVIEW source code automatically will be Open Source if NI folds up or decides to shut down LabVIEW development. Selling the rights to it to some other party who declares its intention to somehow continue its development, is almost certainly enough to fulfill all obligations of such an escrow agreement.

As Brian (Hooovahh) already mentioned, open sourcing LabVIEW is not an easy task. Aside from old legacy "crimes" committed in the early days of LabVIEW development (and often technically unavoidable since there simply weren't better technologies available, LabVIEW pushing the economically available hardware to its limits, and/or LabVIEW's source code requirements pushing the limits of what C compilers could do back then), there are also things that are likely simply not open sourceable for several legal reason. Reworking the source code to be able to publish it as open source would be costing a significant effort. And who is going to foot that bill, after NI is gone or decided to stop LabVIEW development?? For instance, I'm pretty sure that almost nothing of the FPGA system would be legally open sourceable, since it is probably encumbered with several NDAs between NI and Xilinx. And there are likely several other parts that under the hood are limited in similar ways.

Even just the effort to investigate what could and what couldn't be open sourced is going to cost some serious legal and engineering man power and with that also real money. Then someone has to remove anything that was identified as being impossible or undesirable to be open sourced and clean up the mess after that. This in itself would be likely a serious project. And then what? Who is going to maintain a project like that? .Net Core only really is successful because Microsoft puts its might behind it. Without Microsoft's active support role it would be still Mono, hopelessly trying to catch up with whatever new thing Microsoft comes up with.

Edited June 11, 2021 by Rolf Kalbermatter

[ShaunR]

1 hour ago, Rolf Kalbermatter said:

Selling the rights to it to some other party who declares its intention to somehow continue its development, is almost certainly enough to fulfill all obligations of such an escrow agreement.

The most likely route is to carve off LabVIEW into a separate subsidiary company to let it sink or swim on it's own merits.

[X___]

Let's start with unlocking all VIs and documenting Math DLLs...

Edited June 11, 2021 by X___

[Rolf Kalbermatter]

1 hour ago, X___ said:

Let's start with unlocking all VIs and documenting Math DLLs...

Why?