Re: Software Pricing (Was: A Letter To The FLEX-ES Community)
In a message dated 10/17/2006 10:09:59 A.M. Central Daylight Time, [EMAIL PROTECTED] writes: As I understand it, if A is 37% cheaper than B, then it costs 63% (100-37) what B costs. I quite clearly stated that I meant a factor of 37 and _NOT_ 37%. I made the distinction so as to avoid this kind of inanity, but it seems someone will always be dumb enough to step forward. Your original post, on 12 OCT 2006, said I've seen an IBM internal analysis of a Websphere Application Server implementation that was 37x cheaper on Intel than on zSeries. That's 37 _TIMES_ - not 37%! You made it abundantly clear that you meant a factor of 37 and not 37%. But you sowed confusion by adding the word cheaper. Your post was followed on 13 OCT 2006 by Matt Simpson's post in which he said Statements like this always confuse me. How can something be 37 times (or 3700%) smaller or cheaper than something else? I posted an explanation, shortly after Simpson's describing his confusion over your choice of words, on the basic mathematics and the wording one must use to communicate a percent change in order not to confuse readers who are well versed in basic mathematics, as are Simpson and I. Simpson was not confusing 37 times with 37%, as you apparently thought. He was questioning how a price can be reduced by more than 100%. Basic math states that the percent change in moving from X to Y is 100*(Y-X)/X, unless X=0. E.g., moving from 10 to 20 is a 100% increase of the beginning value of 10, from 10 to 0 is a 100% decrease, and moving from 10 to -40 is a 500% decrease of the original value. This works fine in abstract math but not always in the real world. I'm quite stunned at the apparent ignorance of basic mathematics shown in the responses. Simpson and I are both stunned whenever anyone unthinkingly throws in words like cheaper or less with a % value greater than 100, indicating the writer has succumbed to the inane, endemic mind-rot sown by advertisers, media hypers, and politicians who, in trying to gain the attention of the reader, claim that something has been reduced in price by more than 100%. These hype-mongers are the ones who are dumb enough not to understand basic mathematics. Simpson's use of 37% was a hypothetical attempt to understand your confusing use of the word cheaper, not an attempt to claim that 37 times is the same as 37%. In your latest post you have cleared up the confusion by saying I HAVE IN MY POSSESSION AN IBM INTERNAL STUDY SHOWING THAT THE COST OF PROCESSING ONE CUSTOMER-SPECIFIC SAMPLE WEBSPHERE APPLICATION SERVER TRANSACTION ON XSERIES IS 1/37TH THE COST OF PROCESSING THE SAME TRANSACTION ON ZSERIES. Thank you for rewording your original statement into a non-confusing and mathematically precise wording to express the basic math involved, which neither Simpson nor I misunderstand. 1/37th the cost is not the same as being 37x cheaper. Engineers, architects, and even corporate accountants are careful to resist dumb and inane wordings when describing percent changes. Bill Fairchild, B.S. Applied Mathematics, 1967 -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Software Pricing (Was: A Letter To The FLEX-ES Community)
Bollocks. I've seen an IBM internal analysis of a Websphere Application Server implementation that was 37x cheaper on Intel than on zSeries. That's 37 _TIMES_ - not 37%! -- Phil Payne http://www.isham-research.co.uk +44 7833 654 800 -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Software Pricing (Was: A Letter To The FLEX-ES Community)
Sorry, I disagree. Mainframe software pricing has been falling, quite precipitously in many cases, over several years. There are a variety of ways that's been true, and a variety of reasons, but it's fact. And the market trends show no sign of abating (personal view), so I expect further decreases. In contrast, non-mainframe software pricing has been soaring. Mainframe software pricing is *not* linear (at least in IBM's case -- can't speak for other vendors). Don't know how that rumor got started. If you look at the Value Unit Exhibits in IBM software announcements it's quite obvious pricing is not linear. Pricing is quite substantially sublinear: each incremental unit has a progressively lower price. The fact much IBM mainframe software is available in smaller quantities (at smaller prices) than non-mainframe software -- WebSphere Message Broker cited as an example -- is indeed a very big deal. Why wouldn't it be? Yes, 3 MSUs of WebSphere Message Broker is productively useful in real customer situations. I try to avoid unreal hypotheticals -- I'm citing an example from recent experience. There are other examples, like WebSphere Process Server and WebSphere Commerce Server to pick two more. (WCS is available on Linux on z.) Re: the situation of smaller z/OS developers: To convince somebody at IBM (way above me) there's a problem (and how to fix it), here's how I'd go about making the argument: 1. Explain why smaller z/OS developers are important. That ought to be fairly easy. 2. Explain what changed for the worse and how much worse (or what didn't change but needs to change, and how). I'm a little puzzled because, over a decent time span anyway, I don't recall z/OS development resources ever being cheap. (When was this mythical those were the days! everyone is talking about? Wasn't it a lot more expensive to write and support code for MVS in, say, 1986?) Is today's price a record low, or is it getting worse? That's an important question, and I honestly don't know the answer. 3. Explain the business impact. A $1,000/month expense for a software company making $1,000,000 per year in profit isn't a bad situation, for example. But reverse those numbers and it's a huge problem. What is the real world impact to individuals, partners, and customers? What would happen (good and bad) if IBM were to make the change? Apologies if all that is obvious, but hopefully it's still useful. - - - - - Timothy Sipples IBM Consulting Enterprise Software Architect Specializing in Software Architectures Related to System z Based in Tokyo, Serving IBM Japan and IBM Asia-Pacific E-Mail: [EMAIL PROTECTED] -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Software Pricing (Was: A Letter To The FLEX-ES Community)
On Mon, 9 Oct 2006 17:40:05 +0900, Timothy Sipples [EMAIL PROTECTED] wrote: Sorry, I disagree. You disagree with what? You didn't quote anything. Mainframe software pricing has been falling, quite precipitously in many cases, over several years. That's a powerful assertion. Back it up with data. Specifically about MVS and related software. The 10% or so adjustment in MSU ratings for the new generations is a tiny drop. There are a variety of ways that's been true, and a variety of reasons, but it's fact. And the market trends show no sign of abating (personal view), so I expect further decreases. Not enough to make a significant difference, IMHO. In contrast, non-mainframe software pricing has been soaring. Really? How about some data? The mainframe is the only place that I know where the software pricing is tied to the compute power of the processor Mainframe software pricing is *not* linear (at least in IBM's case -- can't speak for other vendors). Don't know how that rumor got started. The phrase I used was nearly linear and I stand by it, but I admit that it is from memory. I have no data available to me. It would seem you do, but you do us no service to make these claims without presenting data to support them. If you look at the Value Unit Exhibits in IBM software announcements it's quite obvious pricing is not linear. Pricing is quite substantially sublinear: each incremental unit has a progressively lower price. There is no such section in the z/OS 1.8 announcement. IBM stopped publishing the price for software in announcements many years ago. Is there a web site where MVS (and related) software pricing is documented? I searched and couldn't find it. I used to use it frequently. The fact much IBM mainframe software is available in smaller quantities (at smaller prices) than non-mainframe software -- WebSphere Message Broker cited as an example -- is indeed a very big deal. Is it a fact? Can you give any numbers to support that assertion? Tom Marchant -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Software Pricing (Was: A Letter To The FLEX-ES Community)
Tim, let me give this a shot. 1. 1. Explain why smaller z/OS developers are important. That ought to be fairly easy. Fairly easy for me to explain to a friendly audience. Very difficult to explain to an endless chain of executives who don't see how small software developers affect their bonus and or objectives. 2. Explain what changed for the worse and how much worse (or what didn't change but needs to change, and how). I'm a little puzzled because, over a decent time span anyway, I don't recall z/OS development resources ever being cheap. (When was this mythical those were the days! everyone is talking about? Wasn't it a lot more expensive to write and support code for MVS in, say, 1986?) Is today's price a record low, or is it getting worse? That's an important question, and I honestly don't know the answer. No longer really being on the business side of MVS software development, I'm not sure -- but yes, the P/390 announcement was the golden day of small-company MVS development. Where this thread started is that it looks like its more complex, more expensive successor is going away. That is a turn for the worse, for the way worse. 3. Explain the business impact. A $1,000/month expense for a software company making $1,000,000 per year in profit isn't a bad situation, for ... Woo-hoo! It must be nice having that perspective on what a small software company is! My company had the #3 product in its mainframe category and the most profit we EVER made in a year was about $250K. Not real useful to talk about company size in terms of profit because profit is affected by so many factors. Better to talk about company size in terms of revenue. A $1MM profit would probably be a company with around $10MM in sales. $10MM in revenue would make you about the 400th largest SW company in the world. (Source: Software Magazine 2005 The Software 500) I'll bet the Dave Salt's of mainframe development would kill for that kind of revenue. No, we're talking about the one-to-three man shops, where the real creative stuff comes from. A prosperous three-man shop might have revenues of $400K-600K; a struggling startup might have revenues that were much, much less, and so yes, $1000/month is a BIG deal. And the point of the thread is, I believe, that if FLEX goes away, the cost will be much, much more than $1000/month. Does this help? Charles -Original Message- From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] On Behalf Of Timothy Sipples Sent: Monday, October 09, 2006 1:40 AM To: IBM-MAIN@BAMA.UA.EDU Subject: Software Pricing (Was: A Letter To The FLEX-ES Community) Sorry, I disagree. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Software Pricing (Was: A Letter To The FLEX-ES Community)
On Mon, 2006-10-09 at 06:24 -0700, Charles Mills wrote: Tim, let me give this a shot. Touché Charles - good post. Shane ... -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html