1 million waveforms per seconds sounds great but to me that's just marketing wank (others have done it before, too). Update rates are way overrated
1 million waveforms per second is about high acquisition rates for low blind time. There is some minimum dead-time between acquisitions which limits it. This is what matters when searching for rare signal characteristics.
High update rates seem to be pretty pointless for searching rare events
https://forum.allaboutcircuits.com/blog/the-truth-about-oscilloscope-waveform-update-rates-and-why-not-to-fall-for-it.1514/
Before writing anything else, I'm sorry we got into argument. And what I write further is not a provocation or anything.
I just want to discuss this, fact based, like you (and also me) like it.
Let me go by the numbers it is easier to be clear and to avoid language barrier.
1. 1 MWfms/sec IS a marketing gimmick, and for more than a few reasons. Most of all, at most timebases it won't be that fast, if you have trigger holdoff it won't be that fast, if you have trigger delay it won't be that fast.
2. Short retrigger time IS NOT a gimmick. Short retrigger time is a positive consequence of them chasing 1 MWPS. That is a real benefit they don't market. In interactive, visual mode scope feels analog like. Also in segmented mode, you don't miss sequential events..
3. If you are trying to characterise some signal that is well defined as how it should look (serial comm bus, clock etc ), something for what you have a clear specification, then staring in the screen for something to happen is definitely NOT best way to capture anomalies and best use of your time. With that kind of signal and based on specification you can devise set of triggers, searches and stats that you let run for few hours (or days, doesn't matter) and when you come back you get stats and your offenders will wait for you neatly stored in history buffer.
He is SOO right about that. Funny thing is, you can do the same thing on Keysight and all other scopes. Heck, little Rigol 1000Z has so many advanced triggers like many of the midrange scopes of the yore (not very good documentation that will explain how to use it though). That kind of workflow is nothing exclusive to LeCroy. It's just since on LeCroy that is prefered workflow, they explain it and endorse it, because it plays to the strengths of their platform.
BUT also, in a pinch you could enable persistence and go to lunch and, when you get back, you take a look at the screen to see if there was a glitch. It will catch it, but my biggest problem with that is that you won't know when it happened, how it happened and have no way to know that. You will just know something did happen, and now you still have to catch the culprit. So you still have to devise detection protocol, set triggers etc.
So that is only partially useful. But it does provide differential diagnosis, meaning that if you don't catch anything with persistence, and you didn't catch anything with a protocol of various triggers, it enhances probability you are glitch free. So IT IS useful for that. By itself NOT much.
4. If you are working on designs where you are scoping (is that a word?) just some nodes in a circuit, or you are designing something, signal will be nothing like clock or something that can be easily explained with few rules. In that case it is all about you and not the scope. Any scope will at that moment have some good and some bad things. Generally, what is useful is good measurements, history(segmented) buffers, search, stats... Some will LIKE fast interaction, some will NEED deep analysis. I like both, Keysight for interactive, Picoscope to get data on PC and then i can do anything I like. Use of brain is not optional here, no universal answer. Workflow will depend both on problem at hand and every individual skill-set, habits, and how people LIKE to work. Some people have no problem whipping up a custom analysis in Mathlab or Labview in just few minutes, and may actually prefer it that way. Some will insist that scope has to have certain analysis built in so they can just use it. Your mileage can vary.
5. Thing about Mr W. is that he doesn't explain why and how he came to the conclusions he proclaims "The Truth". It is not that he's wrong about the topic (he is not right about everything but mostly he makes a good point.), but also didn't provide not even a thought process how he came to his conclusions, much less concrete examples to make his case stick. He just say "this is stupid and waste of time" and "this is solution for all problems". Long time ago when I read his original posts here, I got curious and did a little research. And I realized that he made a lot of good points but his causalities were not always correct. He was basing his assumptions on marketing material and positions of manufacturers. So he concluded that short memory / high WUR scopes (like Keysight) are inferior to long memory / low WUR (like LeCroy) scope, because on LeCroy you CAN use triggers to accomplish so much more than on Keysight with high WUR and persistence. Which was misdirection in terms.
Correct conclusion is that
staring at the screen, looking for a glitch is INFERIOR and LESS productive way of doing it, as opposed to setting triggers, segmented memory and automating search for anomalies. Problem is that Keysight can do that to, the trigger way, for most of the part. Lecroy scopes in that price range and class (Wavesurfers) have none of the advanced features of high end Lecroy machines that makes them so awesome. On lower end Lecroy triggers are not more intelligent than ones on Keysight (or Rigol to that matter), and Wavescan is just fancy name for a search.
Fact that Keysight trumpets about WUR and catching glitches with persistence doesn't mean I have to use it that way. So I use Keysight like he uses LeCroy, with prudent choice of triggers and segmented memory, using pretty advanced search capabilities to drill down further. And it works, well. He was right about it, thanks a lot for that. But his dogma was you need a LeCroy to do so. You don't. On high end Lecroy with advanced analysis options enabled, you will do it faster, better, easier. But that workflow works well even on lowly Rigol, and it works well.
Actually, exactly that type of workflow enables scope that doesn't have much memory (like Keysight) to not show that. Long memory is not that important in that case because you only capture what is of interest and ignore hours of signal you don't care about. So it actually complements Keysight well, hiding it's imperfections, instead of making it unusable.
We could go on for years like this. Everybody will protect their choices, their cognitive biases, their way.
Fact is that a good, working LeCroy Waverunner in a good shape is a hell of a scope. If you can enable advanced analysis options, you will have a scope that you can develop stuff NASA would be proud of. OTOH, some people only need scope for servicing stuff, and they want fast and familiar.
All the best,