| About InfoWorld : Advertise : Subscribe : Contact Us : Awards : Events : Store |
 |
||||
|
||||
 |
||||
 |
|
|||
 |
||||
|
|
|
||
|
||||||||||||||||||||||||||||||||||||||
 COLUMN TO THE EDITOR March 23, 2001 See correction below
I AM GETTING tired of reading articles like the one by Mark Leon, "A Union of their own?" (March 12), that treat unions as if they were an infestation of rats. (Oh my gosh! We have unions!) Unions start at companies because the employees feel their concerns are not being addressed by management. When employees feel like they have no power, they organize (i.e., form unions) to help them regain some influence over management decisions. The solution for management is not to fight unions, but to solve the problems that cause employees to want unions. If employees are working 12-hour to 14-hour days without what they consider to be "just compensation," then either cut down on their hours, pay them more, or give them some other bonus to make up for the long hours. When management begins to think of their employees as simply cogs in their great profit machine, then unions are bound to happen. Union membership comes at a price, and most employees understand that. The job of management is to make the working conditions, pay rates, health plans, etc., good enough so that at least 51 percent of their employees will feel that getting a union is just not worth the trouble. Douglas Mach, Hartselle, Ala. A good sign I WAS PLEASED by the positive portrayal P.J. Connolly gave IBM's AIX 5L (see "Linux and Itanium alter Sun's gravity: IBM proves its commitment to Linux with AIX 5L offering," Feb. 26). It is a good sign that IBM has made the proper value proposition for the next-generation AIX. But there was a comment in the "pros and cons" section of the bottom line box that I must take issue with. Connolly knocks IBM hardware as "still lagging behind" both Hewlett-Packard and Sun in terms of numbers of CPUs supported, and as lagging behind HP in the total amount of memory supported. The last time I checked, the IBM servers which run AIX 5L support up to 96GB of real memory. Although HP has shown a NUMA (nonuniform memory access) server design on paper that will exceed this, I doubt that one has been installed anywhere. And I doubt that there are many workload mixes that can exploit even 96GB. As to the number of CPUs supported, it is easy to write support into the OS for additional CPUs. What is much more challenging (as Sun's and HP's performance and scalability demonstrate) is to get reasonable throughput and capacity from those incremental resources. IBM has clearly demonstrated superior performance and throughput scalability with fewer CPUs than either Sun or HP. Chub Varga, Cranford, N.J. HP is a gutless company? COMRADE PETRELEY THINKS that "HP is a gutless company" for not offering up the source code to their OpenMail product (see The Open Source, March 12). How many years did HP spend developing this software and at what cost? Mr. Petreley would toss aside such trivial concerns as contracts governing non-HP intellectual property in order to stick another barb in that devil Microsoft. I doubt that businesses around the world who rely on OpenMail would like the source code to their e-mail systems suddenly made available to competitors and criminals. Please leave HP alone, Mr. Petreley. Few companies in any industry can match HP's record of long-term rationality. Don't whine about a nice boy who refuses to hand over his lunch money. Take your "open-sourcers," go try to build your own business, and quit bumming cigarettes from the guys who have day jobs. Scott Schad, Tulsa, Okla. Industry-standard restrictions THE COVERAGE OVER the past few weeks in Robert X. Cringely's column about benchmark publication restrictions really misses the point about why these restrictions are industry-standard practice (see Notes From the Field). It is not worthwhile here to argue the specifics of the situation described in the Cringely columns; instead, I would like to describe Microsoft's overall position as straightforwardly as possible, and invite your readership to provide their views on this topic. First off, let me state that the "no benchmark publication" clause in our license agreements is neither unique to Microsoft nor meant to restrict any customer-to-customer communication of application performance seen in individual environments. Simply put, these restrictions are there to ensure that published data is accurate. There's no more to it than that. We have approved accurate reports regardless of what they say about our performance, and we will continue to do so. Database benchmarking is a complex process, and the industry has built a number of standardized tests where hardware, software, and applications vendors work together to create well-measured, rigorously audited results. These standard benchmarks are only published once all the parties are satisfied that they have achieved their best performance. In other circumstances, software labs will occasionally construct one-off benchmarks to illustrate a particular point, and our benchmark publication clause merely causes the lab to involve Microsoft in the process early enough to ensure a thorough examination of the testing configuration and methodology. In the most extreme scenario, the benchmark publication restrictions prevent disreputable competitors from fabricating bogus tests and broadcasting results that purport to show their superiority in some aspect of product functionality. All these procedures are meant to ensure that a fair, accurate, and truthful representation of product performance is communicated to the public at large. For years Microsoft has published industry-standard database benchmark results including three current TPC scenarios, three separate SAP modules, plus line-of-business applications from Peoplesoft, JD Edwards, Baan, Great Plains, Onyx, Pivotal, and others. Our results have sometimes been far from the top, but we were paying our dues in a marketplace new to us. Now that we are obtaining top marks in most of these benchmarks, we can proudly say that we earned our position through smart development and hard work. Our objective in reviewing benchmark results is simply to ensure that results reflect the true capabilities of our products. We are interested in hearing from your readers on the topic of benchmarking and vendor restrictions. We've set up an e-mail account to obtain feedback on this topic: www.bnchfdbk@microsoft.com. We invite your readers to tell us what they think about benchmarking processes and license restrictions. Steve Murchie, Group Product Manager, SQL Server Microsoft, Redmond, Wash. Got spam? CARLTON VOGT'S RECENT column "Unsolicited e-mail: What the devil is wrong with it?" (Ethics Matters, March 6) misses two important points. Most unsolicited commercial e-mail is carefully crafted so that the recipient cannot easily reply. There is no effective way for me to contact someone to say, "I really do not wish to receive this." As for those that do allow someone to reply or give a "remove" address, the typical spammer often sells the list of "removes" as a "qualified list." It seems morally and ethically correct to honor remove requests for what they are, but instead it is the best way to receive more unsolicited commercial e-mail (UCE). The second issue is how UCE is sent. Read the headers -- they virtually always lead to a server that is not properly secured. Sure, the server owners should do their best to make the server "safe" from this kind of abuse, but an active spammer is still stealing resources. Lynn Taylor, Laguna Beach, Calif. CARLTON VOGT'S PREMISE is flawed in his analogy that bulk mail supporting postage prices is like unsolicited commercial e-mail. Bulk mail is paid to the Postal Service by the sender, thereby relieving the delivery service of the additional volume represented by bulk mail. UCE on the other hand does not perform the same function vis-à-vis subsidizing other Internet services. The spammer pays only its own ISP for the connection to the Internet. A much better analogy would have been receiving postage-due solicitations or accepting collect telemarketing calls. Imagine what would happen if every business in the country sent UCE. Our e-mail boxes would become unuseable, your ISP would be unable to keep up with the load, and e-mail would become a useless mechanism for communication. There is a reason for ISPs to include a no UCE clause in their terms of service, and that is to keep e-mail a functional medium, not one rendered useless by unfettered UCE. Mark Lange, Tucson, Ariz. In our March 26 Letters to the Editor we misprinted the e-mail address of Microsoft's benchmark feedback mailbox as supplied by Steve Murchie. The correct address is bnchfdbk@microsoft.com.  RELATED SUBJECTS  E-CommerceSecuritySoftware DevelopmentSPONSORED WHITE PAPERS 
SPONSORED LINKS
|
||||||||||||||||||||||||||||||||||||||
|
||||||||||
|
||||||||||
