Embedded Systems Programming drops the ball.

Hello,

I read with great interest the above comments. CMX, too, was very disappointed with the published survey. The CMX-RTX RTOS has been available since 1990, has over 4,000 user companies, supports more processor cores than any other RTOS vendor (approximately 50) and CMX has been involved with the ESP magazine and its trade shows for many, many years. How CMX could be overlooked in this survey is baffling.

JR Rodrigues CMX Systems, > I'm not usually one to gripe in public but the latest issue of Embedded

How many full page advertisements have you bought recently? :)

I have zero trust of "surveys" or "industry averages" published in commercial magazines.

I was wondering about that line "and CMX has been involved with the ESP magazine and its trade shows for many, many years" and just how it was supposed to affect this product's placement in this OS Survey article. I've got nothing against ESP, I used to subscribe in the late '90's and wish I'd kept it up, but isn't it about time for "open source" magazines, user-edited much like Wikipedia? Or why not write an "embeded OS" article on wikipedia? I looked around a little, there's not much on embedded, there's this entry that looks like it could use an addition, even if it's by someone connected to the company:

To repeat a six million dollar cliche, "We have the technology."

-----

The first part of results of that survey, concerning processor choices, also had a number of serious flaws (IMHO), which I think mostly come down to failing to define their terms. For example, there was a section showing the mixture of single-processor systems and multi-processor systems, with no indication as to how they defined this. It's easy to define this on a PC or a server, but is a data acquisition system with a single central board and hundreds of networked intelligent sensors a single-processor system or a massively multi-processing system? Without knowing that, the data was meaningless. (Jim Turley replied to my emailed questions with some explanation, but still failed to clarify that issue.) There was also a category for the spread of processor size, without the size being defined, so as usual people will have used conflicting and frequently absurd definitions of size (there is even a category for 10/12/14-bit processors !).

I guess you only ever expect to get a rough idea from these sorts of surveys, and not concrete answers. They are heavily biased by the surveyor's pre-supposed ideas, which seldom correspond with the reader's pre-supposed ideas.

David

In article , snipped-for-privacy@larwe.com writes

Why? they are as accurate as any other. Also some of the ones I have seen are very good because the magazine in question could get real information that is not always released publicly. Also the magazines are usually well placed to see through the marketing spin (if any) on the information they receive. From experience these magazine editors get a lot of access and information that ordinary mortals do not. This means that they are usually better placed to produce these surveys.

On the other hand much of the open source stuff is driven by an almost religious zeal that can distort things. This makes many companies very wary of dealing with them. If they are not "on the inside" they will not get the accurate information to produce these surveys.

I think the implication was that: "they know who we are and where we are" so one would expect that CMX would be in a survey of OS's. I don't think the implication was that they were looking to "buy" a good review.

This make one wonder what the selection criteria was if two of the common (widely known) OS's were missed out.

No.

Edited by who? a single person has to do the editing. This takes time and effort. If everyone is doing everything for free where does one earn money?

The one way of getting things done for less cost is to out source to the Far East. This helps the open source community as it gives them much more time to work on open source projects as they have no full time work. They will of course have an active "customer" base of the working engineers in the far east.... However it will be one way traffic.

...that we are putting in to open source and shipping out the competition...

Why? I have a similar distrust of open source stuff having seen some of the Zelots and strange mindsets that are involved.

Well, Chris, first I'll mention a logical flaw in your comment here -- if it is your intent to disprove Lewin's well-advised caution -- then I'll get to my comment towards the meat of the matter.

Lewin's comment is entirely consistent with yours, above. You go on about "Also..." after that, but it these two comments, from you and from Lewin, that remain entirely consistent. Yours doesn't dispel his in any way.

But to the meat: I've personally seen the purchase of opinion pieces of staff writers at well-trusted, widely circulated newspapers and also, similarly, those staff writers of a national magazine or two -- without them disclosing anything about the financial relationship, so that readers would take the impression that the reviews were "above board." There _is no_ good reason to inherently trust any review and there _is_ good reason to always ask yourself _if_ such motivations _may_ apply when reading them.

Never assume that business isn't business. You may imagine that an author "would never sully his or her name in such a way because they have a reputation in the industry," or that a magazine or newspaper wouldn't ever put themselves in such a risky position, but that may or may not be true and unless you can otherwise justify your assumption of honesty, you should remain circumspect. Respected-name reviews in the software review industry have been (and probably still are) bought and sold. Not always, not mostly, but sometimes and often enough for caution.

As Lewin puts it, I trust no review to be entirely above board unless I write it and personally know I didn't receive any "payola." And even then, I'll still worry a lot.

Jon

As Jonathan pointed out, I believe that was actually my point. I'm not saying they are all flawed. I'm saying that I don't know, but that I _am_ aware of commercial bias (look at Microsoft-sponsored magazines).

Maybe the reason XYZ product isn't in the survey is because nobody at the magazine knew about it. Maybe it's because they only found out about it by feedback during the survey - at which time it was too late to re-print, re-distribute and re-collate the results. Or maybe it's because a big advertiser said "Don't mention XYZ or we'll pull our business".

A survey like this doesn't tell you what was left out or why; it's noise, not information.

If they don't mention ucOS-II, this is probably not because they haven't heard of its efficiency, its broad support of a quantity of devices and its safety level(DO-178B level B).

Probably because it is a threat to expensive rtos that want to dominate the market. Fair competition sometimes means biased competition for some.

I wonder if it would be possible to make some kind of survey in this NG, by pointing out the features of rtos effectively used and their applications. I know advices have been given previously on some OS but it would be interesting to have the opinion of users independently of vendors or magazines.

In article , Jonathan Kirwan writes

I take your point. However this is true of virtually any survey.

I think this is a reasonable stance.

A lot of it is down to copy deadlines. I have seen surveys that miss out some major players because there was no information supplied to the magazine in time. The authors write reviews but often need the company whos item is under review to supply information/ equipment. Otherwise there is just not time to go and get it all.

In article , snipped-for-privacy@larwe.com writes

I take your point. There are some "sponsored" magazines (commercial or otherwise) who do have a particular slant.

This is the usual reason.

This is (I hope) less common.

Some do tell you why some things were left out but it is impossible to list (do a survey? :-) of everything not included.

Between the devil and the deep blue sea.

My point was that it is any survey not just the commercial ones.

In article , Lanarcam writes

Was that a comment or an advert :-) I am sure they have hear of it the question is why was it (and CMX) not in the survey?

This is what I mean about bias. We are discussing CMX AND ucOS2 one is a commercial OS and the other is a commercial OS. Neither is expensive compared to some.

You did not include CMX in your comment. Why not? Biased? Paid to comment on one and not the other? Working for one of them? Getting free samples? Walking out with he son/daughter/goat of the owner? :-)

BTW I have nothing for or against either. Though I believe both to be sound OS.

This is no more independent than any other survey. most (99%) of commercial companies in the business monitor this NG and most of the post using non company email addresses. However probably only 1% of the embedded community use this NG. So any survey here is also going to be biased.

A simple comment from the study of the excellent book that goes with it and the comparisons with other in house or commercial rtos. I have no commercial link with the author although I had excellent contacts at least for one tender. Besides, I also have good contacts with the local distributor. The fact that the source code is freely available and that it can be used free for educational purposes tells also a lot.

I don' know CMX and I am not biased against it of course.

I don't know his daughter but if you have insights, do tell me:)

I didn't have a lot of succes so far, so let's assume you are right, have you any better idea?

... snip ...

Not all that long ago magazine policy was to publish a notice: "We are doing a survey on foo; please submit bar on your foos by blah for inclusion". Then the final article might have a notice "Data on X and Y was received too late for inclusion, and we have mentioned fum because its foo is well known".

In article , Lanarcam writes

Unfortunately no.