Thread regarding SAS Institute layoffs

Compiling C to Safe Rust : https://arxiv.org/abs/2412.15042

The Microsoft partnership should come in handy in obtaining access to the expertise required for the conversion.

Good news! Expect a cybersecurity regulation rollback and elimination of onerus security requirements. Memory safety will no longer be a problem:

https://www.wired.com/story/trump-administration-cybersecurity-policy-reversals/

Long live C/C++.

@7yko+1v7RIXtX

Could be that this site is just glitchy but it appears that the posts you reference as insightful have been removed. Apparently criticisms of Viya are censored when such contain too many implementation details … even when these are derivable from public domain sources in most cases published by SAS.

They should rewrite everything in Flowgorithm.

I’ll just leave this right here:

https://thenewstack.io/feds-critical-software-must-drop-c-c-by-2026-or-face-risk/

Feds: Critical Software Must Drop C/C++ by 2026 or Face Risk
This is the government's strongest stance yet on software security, which puts manufacturers on notice: fix dangerous coding practices or risk being labeled as negligent.

"How are sales of VA?"

Not good enough to move the revenue needle???

Likewise, sales numbers for VA will be the same as for Viya, because they’re bundled.

A search for jobs using “SAS Visual Analytics” returns exactly one, and it is consistent with the statistics-related job postings described by @9wzf+1v7RIXtX:

“3+ years of developing visual analytics reports in SAS Visual Analytics, SAS Visual Investigator, SAS Studio, Tableau, QlikView or similar technology.”

How are sales of VA?

Take a Statistics class and argue about which program you complete your analyses in. The Professor doesn't care. They simply say "Those are technologies. We want you to understand the fundamentals. You can transfer the fundamentals to any technology."

That being said, when I look at Statistics-related job postings, the postings have "skilled in SAS, R, Python, etc." My interpretation is that the organization doesn't care what you used, they have their platform and want you to know the basic dance steps. Unfortuantely, they don't reveal which platform they use.

The count of Viya-related jobs is interesting and specific. Maybe it shows something, maybe it doesn't. There's not enough information about it.

Customers can't purchase "Viya only", because Viya includes V9.

Therefore, numbers can't show "Viya only".

They show both products, and you can't tell which one the customers are using.

@9fhd+1v7RIXtX “ Some people can't be helped.”

Agreed. Check your mirror.

@9fhd+1v7RIXtX Imdid answer the question more than once. You just don’t like the answer.

And no I am not sharing internal confidential numbers here. If you work at SAS look it up.

@9mev+1v7RIXtX

You didn't answer the question.

And the numbers you claim to have -- you never show them.

Some people can't be helped.

“ so the “Viya only” numbers you describe cannot exist.”

LOL. Yeah ok.

@8ljq+1v7RIXtX

If Viya and V9 are sold as a bundle, then every Viya sale includes a V9 sale, so the “Viya only” numbers you describe cannot exist.

Do you claim that they are not bundled?

“ Cutting a pile of BS in half still leaves a pile of BS”

I like your analogy but technically cutting a pile of BS in half leaves you with two piles of BS.

@8voy+1v7RIXtX Numbers I’m looking at are Viya only. But keep reading LinkedIn.

@8faw+1v7RIXtX

My understanding is that Viya is sold with V9. If so, the “actual numbers” you’re looking at don’t show whether customers are using Viya or V9.

Hiring numbers show hundreds of customers strongly committed to V9 — enough to hire for that skill. Only about ten are hiring for Viya.

@8apb+1v7RIXtX Oh well if you heard it from a friend….

I’ll continue to go based on actual numbers that I’m looking at as I type this. Seems slightly better than from anonymous site guy who heard it from a trusted source.

But, Workbench.

@7olj+1v7RIXtX
How would you know if you don't even know which office I'm talking about? From 7 to 8 figure sums to virtually zero, from a source I trust. I believe it based on what I knew of the pipeline when I left.
And the number of job postings for a given technology is a pretty good proxy for product adoption. It's the best signal we've got due to SAS's unwillingness to come clean on the actual numbers which they used to disclose when times were good.

“ like someone said 1 customers to 2 is 100% growth”

And if those were the actual numbers that would be concerning. But those aren’t anywhere near the actual numbers so…

Cutting a pile of BS in half still leaves a pile of BS.

The numbers are available if you work at SAS. Look them up if you work there.
Anyone who doesn’t work there or doesn’t care about using meaningful data keep on making sh-t up.

Obviously the company won't say other than mention "growth" (like someone said 1 customers to 2 is 100% growth). If there were very large numbers involved, maybe more would be said, but not necessarily. Related job posting figures relative to other technologies do give some objective metrics of related activity.

"I’ll take actual sales numbers thank you very much."

So, go ahead and post them.

Returning to the OP.

A significant issue in a Viya CAS, etc. C/C++ conversion to Rust Is the current reliance on the homegrown proprietary SAS TK libraries. There are many complex TK extensions supporting various services specific to SAS data access + analytics along with all sorts of common DSA utilized throughout the current Viya codebase. There are a few hundred thousand lines of this common functionality TK extensions code that has bern hand optimized over the past 25 years.

Would this code also be converted to Rust? Doing so would be very tedious and fraught with risk. Effectively testing the result would require seasoned pros with senior level development expertise, historical knowledge of TK and in many cases heavy DSA and math skills.

At this juncture in time only an ambitious fool with TONS of money to burn would consider green-lighting such a conversion.

@7vfz+1v7RIXtX It sure what office you are talking about but no sales in 2 years is not even remotely the case.

I’m not saying LinkedIn is messing with the numbers. I’m saying that is a ridiculous way to try measuring the success of a product. I’ll take actual sales numbers thank you very much.

• 4 needs a little more help, bring on the VRBP!

Can any doubt that this is the plan?

Spoken like a person trying to establish a buyer’s understanding.

Before we got into this discussion about whether numbers matter (hint: they do) I enjoyed the insightful comments from @oip+1v7RIXtX, @4cdh+1v7RIXtX, @5nxn+1v7RIXtX, @5eno+1v7RIXtX, and @5yle+1v7RIXtX.

I learned that:

1. Viya has architectural flaws, to be expected in any product that is nearly ten years old. Some of them may be attributed to a lack of investment in cloud expertise.

2. Rewriting in Rust would fix some flaws, but would not likely increase sales — unless the user experience is also improved.

3. Architectural flaws, particularly memory footprint and performance, increase Viya’s price. This is serious, because Viya competes against open source software including Spark, just like SAS competes against R and Python.

SAS statements and actions are consistent, not with a plan to rewrite and improve Viya, but with a plan to sell the company. If SAS is for sale, they will never make any more huge investments. Rather, they would:

1. Claim that “the user base for Viya is growing” -- without providing numbers, or mentioning that it's growing from a low base.

2. Announce AI projects — essential hype to sell a tech company — even though these projects also compete against open source software.

3. Break up business units into subsidiaries, to make a sale more attractive.

4. Reduce headcount sufficiently to ensure profitability.

5. Announce a plan to IPO.

6. Hope that all these statements attract a sweet offer from a private buyer; but if not, prepare to IPO as a guaranteed exit.

Can any doubt that this is the plan?

As someone who used to work at SAS until earlier this year, and job searched for a role utilizing Python, I agree the jobs out there utilizing SAS are miniscule! Regardless of the official sales numbers, no one can argue against the fact that SAS is on the decline.

@7tzd+1v7RIXtX
I no longer work at SAS so can't look up the latest sales reports/dashboards but I did recently catch up with a ex-colleague who's still there and they told me that in their office, they hadn't sold a single license in over 2 years. When I was there 4 years ago, this particular office had been selling 7 to 8 figure sums every year for many years.

And BTW, what's unconvincing about the LinkedIn numbers? Go check for yourself, it's not hard. Do you think they're fiddling with the search results to make SAS look bad or something? 🤣

“ No one on this thread has produced any numbers to defend Viya; the best they can do is to hack the votes.”

Yeah the LinkedIn numbers are real convincing.

And I’m not even saying it is doing well. But the numbers produced on here to say it is completely failed are laughable.

If you want numbers to “defend/attack” it and you work at SAS then I suggest you read the sales reports or dashboard. Anything else and you are just making sh-t up.

“ presumably by some complete loser who thinks that's rather clever and has nothing more constructive to do.”

As opposed to constructive whining on anonymous internet site? Someone using their skills for something silly is still more constructive than anonymous whining. Be more clever.

Ugh oh gramps is probably mad now.

LinkedIn search LOL. It is funny every time you say it.

‘One day someone will write a PhD on the decline of SAS and the role Viya played, exploring concepts like "sunk cost fallacy" and "founder's syndrome"’

Broken record. One day…

@6uzs+1v7RIXtX

A LinkedIn job search with quotes returns 9 “SAS” “Viya”, 546 “SAS”, and 4,817 “Python”. This still supports your statement that “Viya cannot be credibly defended.”

No one on this thread has produced any numbers to defend Viya; the best they can do is to hack the votes.

When the Ph.D. thesis is written, it will have to reference “The Innovator’s Dilemma”.

@6kue+1v7RIXtX
The downvotes have obviously been automated...presumably by some complete loser who thinks that's rather clever and has nothing more constructive to do.

At this point, Viya cannot be credibly defended.

The platform was released almost a decade ago, yet LinkedIn jobs shows a grand total of just 10 jobs asking for it!

SAS (presumably mostly V9) clocks in at 887 right now.
Python 341,947

What this tells us is that SAS is dying, and Viya was never really alive in the first place.

One day someone will write a PhD on the decline of SAS and the role Viya played, exploring concepts like "sunk cost fallacy" and "founder's syndrome"...

“Let them defend Viya with logical arguments in comments if they want to be taken seriously.”

There is not arguing with inflexibility. Pointless exercise in futility. I’ll just continue to mock people making sh-t up. Kind of speaks for itself.