Syncronys' SoftRAM Scam: What took us so long?

Happy New Year everyone. I hope you had a quiet and relaxing holiday and hope you did too. This week, I look at how our industry tests products and tracks breaking news about these tests using print and on-line resources. Specifically, the SoftRAM Scam.

SoftRAM claims to add memory without the need for buying the actual chips for both Windows 3.1 and 95. The product is at the center of a great deal of controversy: a variety of tests by the trade magazines have shown various things good and bad. Egghead has pulled the product from its shelves. The vendor, Syncronys Systems, at first claimed SoftRAM just had a few bugs and will everyone calm down, then supposedly pulled the product off the market. Meanwhile, the Feds are investigating, and it has become a best- selling product.

I don't really have much experience with these products, although on the Mac side I use RAM Doubler and it truly works (but that has more to do with the fact that programs there don't share RAM as they do on the Window's side). And, having just purchased about $1,500 worth of RAM chips (at about $50/megabyte), I wanted to find out whether that money was really worthwhile or not.

I decided to go on-line and see if I could get to the bottom of this, and came to the conclusion that the print trade pubs have done our readers a tremendous disservice. The best reporting is happening on-line: the information is there sooner and from international sources to boot, and I found a single location that has just about all the sources you'll need to make up your own mind.

To be fair, Bill Machrone first raised some questions about SoftRAM in his September 4th column in PC Week, but the column was devoted to overall problems he had with getting a new machine to run Windows. The first story to focus specifically on anything critical on SoftRAM ran in early-September in the Houston Chronicle with a column by Dwight Silverman entitled, "Test run of RAM doubler is inconclusive." Silverman published several follow-ups with even more skeptical tests.

Another story ran in mid-September in the Sydney (Australia) Morning Herald. The article, "SoftRAM doubles your data space" was written by reviewer Phil Campbell. He started off positive in his article, but then got down to the key point:

"Windows95, on the other hand, has already fixed the System Resource problem - and to be honest, it's hard to spot any difference at all with SoftRAM switched in."

Moving forward in time, the next reference on-line that I could find was in a forum for Win95 that is on Plymouth Commercial Internet Exchange's site. Posters on this system began back then to question the ability of the product to do what it says. I also found, via Computer Select, a critical story in a trade pub called Info Canada that ran in October as well.

So where is our fabled newshounds from Ziff, CMP, and IDG? Not until November is anything written on this subject. This is from a piece called "SoftRAM 95 Does Not Compress RAM In PC Magazine Lab Tests".

"PC Magazine Lab tests indicate Syncronys Softcorp's top-selling SoftRAM95 product ... does not compress memory or increase systems resources under Microsoft Windows 3.x.

PC Mag did additional tests when challenged by the vendor a month later, with the same results. Other articles popped up in Computer Reseller News and elsewhere in November and December. Some mags had printed favorable reviews, then later retracted them when they retested SoftRAM. Some had trouble distinguishing Win 3.1 from 95 issues.

However, on-line is the place to be for tracking this issue. Two places stand out: first, c|net's service carried in early November " SoftRAM 95: the RAM doubler that doesn't", a review by Tinoo Singh:

"SoftRAM provided no benefits whatsoever. Our test systems reported no change in the amount of memory installed after we loaded SoftRAM, and again performance did not improve at all."

Another source with some good reporting is Fa mily PC's Saavy Consumer column by Deborah Branscum. Curiously, the magazine will run the story in their February issue, yet posted the piece on the Ziff web site in late December.

But the best source for tracking this story is O'Reilly's Windows Center web site, and more specifically Andrew Shulman's analysis. Andy probably knows more about the inner workings of all things Windows than most Microsoft developers. He has links to the above stories and more technical details about what SoftRAM does -- and more importantly doesn't -- do. For example, he reports on his own disassembly of the code in SoftRAM to see exactly what is going on. (None of the other sources cited here did that, to the best of my reading.)

"That one can successfully market a product to 650,000 users, with a small number of returns, and receive (until recently) generally favorable press coverage, is a fascinating commentary on the state of the software market today."

Andy gets a Be.Here.Now Award for keeping on top of this story, and going the extra mile in adding a feature that allows you to be informed via email when he updates his web page. That brings breaking news to a new level of service.

So, some thoughts:

  1. Given that the web can track things in real time, I applaud Family PC in putting up Branscum's article now, even though it won't run in the paper edition for another month. And PC Magazine, operating under similar lead times, also has posted news and test results in advance of when they run in the print edition. However, neither can compare with the level of effort, links, and original information sources on all sides of this story with Shulman's page. (I found Andy's page thru Lycos, truth be told.) Shameful.
  2. At the beginning of January, I could still order the software from Software.Net, even though the product had been pulled off the market. You can get it at other Net-based "malls" as well. Ironically, Software.Net offers a link to CMP's Techweb site, where you can read the negative reviews of the product before you order. Shameful, and they get the Web.Feet award for that blooper. They should be more responsible and put up links to some of the above sites so that their customers can make up their own minds.
  3. Why did it take PC Mag and others so long to test a product that has been shipping for several months? Why did other magazines review this product favorably? Why can't we have good reporting in our trade rags on what goes on in Windows 3.1 and 95 with regards to memory and system resource usage?
  4. All of this doesn't bode well for maintaining the public trust that we in the computer trades like to say we have with our readers. Given that the "computer rich," or "enterprise buyers," or whatever they are called these days can get better information on line, get it sooner, and get it straight from the original sources (Shulman's page has links to the statements made by the vendor executives for example), why should they continue to read the print pubs?
  5. If you think that computers are still an American invention, it is time to start looking off-shore and using the net to get better informed. I find it ironic that this story broke first in an Australian daily in September and had a technical review by both Canadian and German computer magazines in October, both long before anyone here in the States wrote about this. Shameful.
  6. How can you, the reader, independently verify trade publications tests? Given the complexity of modern operating systems, the multiplicity of configurations, and the sheer quantity of variables, I agree that it is getting harder and harder to test computer products in any meaningful fashion. And having just invested in upgrading my own test lab myself, I am sympathetic to how hard it is to maintain a state-of-the-art facility.
So the bottom line: if you are running Windows 95, buy those chips like me. SoftRAM, at least in its current incarnation, doesn't work. If you can't afford the RAM, stick with Windows 3.1 and look at RAM Doubler or Quarterdeck's software, which most testers seem to think work.

David Strom
email me
Web Informant site
An archive of my previous essays may be found here