Letters From the end of the month, CNET's Downloads.com page will charge every author $99 for listing the software. Non-payers will see their listings and software removed.

You don't like this, and you don't agree when we argued that CNET had some justification for attempting to cover its bandwidth expenses.

Here's what you say.

Re: CNET intros listing fees for Download.com

Many thanks for bringing this to my attention. This is *dreadful* news for the free software industry. Many authors, like me have no effective way of receiving payments for software - I think that my software would still be successful if there was a decent micro payments system that I could use to charge a couple quid - even one pound for a download of my software.

With Download.com starting to charge (I see their side to the story perfectly), I would not be able to afford the amounts they'll want to list my free software. I receive tens of thousands of downloads of my software from Download.com which I would lose if they started charging.

Dominic Cleal

---

I'm quite surprised by this quote from you:

"We await news of the paid-for conversion rate with great interest, but the proposition makes sense to us. Why shouldn't CNET make a turn on this worthy
service?" [should that be "make a return" ?]

Yes, I can see that the bandwidth costs could be quite high for download.com,
but one wonders if they wouldn't be better charging the *downloaders* (i.e. you and us) a nominal fee - e.g. $10 a year - for downloading, since we're the ones hogging the bandwidth. Disk space is dirt cheap nowadays [so 1TB of disk is not a huge investment and can't be used as a CNet excuse to charge the authors of the listed software], but bandwidth certainly isn't.

Even more surprising about your article is the complete failure to comment on what it will do for the freeware/shareware authoring community - taken to its extremes, if mozilla.org doesn't pay the CNet fee, it could see Mozilla pulled completely from download.com, which would be ludicrous to you and me [it's now probably the best Web browser out there].

Let's also not forget the one-man freeware/shareware operation - they might not be able to afford a download.com listing so bang goes one of the main distribution avenues for them.

How about some more comments on this CNET charging strategy - it seems like you've accepted it without question !

Richard K. Lloyd

---

You say " ...the proposition makes sense to us. Why shouldn't CNET make a turn on
this worthy service?"

That's all very well for shareware and for commercial software, but what about those of us who have written free or public domain software? OK, we're not exactly Mahatma Gandhi, but we are providing a public benefit and we don't earn any money from the stuff we write.

It's fair enough that CNET charges for paid-for software, but why can't they continue to distribute free software for free?

Nick Rozanski.

---

I do really have to say, that it is a sad day when download.com starts charging authors to be listed on a site. I am terribly surprised that you guys of all people would sound so favourably for something of this nature. Yes, I realize we're in an IT slump, advertisement dollar slump...etc. However, One of download.com's real strengths has always been the fact that its tool was filled with valuable and useful free or almost free software.

The struggling shareware author who is likely coding this software as a hobby likely doesn't have the money to fork out the cash for "listing" their items; and what about all that great freeware.

The end of an era.

Jeffrey D Stark

---

"We await news of the paid-for conversion rate with great interest, but the proposition makes sense to us. Why shouldn't CNET make a turn on this worthy service?"

The only people this proposition makes sense to are the CNET fatcats.

As a hobby shareware developer I've spent hundreds of hours developing shareware programs, it seems crazy for me to have to pay $99 every time I release a new version of one of my products, just so someone at CNET can spend 5 minutes doing what could be done automatically.

What about all the quality freeware products out there? Exactly why should freeware developers have to pay $99 every time they release a new version, just so that they can get a bit of exposure for the programs they've spent hundreds or thousands of hours developing?

There's nothing worthy about CNET charging for listing software - listings are their lifeblood, without wares to list they are gonna have some pretty crappy download sites. The process of submission can be done fully automatically, and is by almost every download site out there (see www.filebasket.com, www.softpile.com etc.), except to
CNET's sites. In fact there is an open, XML-based standard for distributing freeware/shareware information called PAD (Portable Application Description) (see http://www.asp-shareware.org/pad/). If the hundreds of small and medium-sized download sites can implement an automated method of allowing authors to upload details of their
products then sure-as-heck CNET can if they would only get the dollar signs out of their eyes for five minutes. Then authors would be happy, downloaders would be happy and CNET could concentrate on making money from advertising revenue and value-added listing services like all the other download sites manage to do successfully.

James Holt

---

Thanks for another interesting article - but I had to ask about the last paragraph: did Ben Silverman really say "enables software vendors to hock their wares" (off to the pawn broker) or did he say "hawk their wares" (as in "software here, get yer s/w while it's fresh").

Though I do like the image of failing e-tailers pawning their products for a last bit of cash.

Eben Gay

Pop-Up Hell

Subject: Annoying bastard ads

What the hell would be wrong with banning those stupid dumbass ads? I hate that crap that scrolls down in front of, across, or between what you're trying to read. The single most bothersome thing about these ads is that fact that they aren't Opera friendly and I CAN'T CLOSE THE LITTLE BASTARDS!

Yeah, that's right. If I go to a site and see one of those buggers start to creep across my screen, I back out and never go there again. Not what they want, is it? I thought the idea was to ATTRACT customers, not alienate them.

Then again, I don't come up with advertising. I just ignore it.

Michael Miller

Pop-Up Hell

Danish watchdog calls for ban on intrusive online ads

Subject: Kasper Larsens pop-up ads/TV-ads comparison.

I am sure that most people would not mind pop-up ads so much if the net was one-way communication like tv. But when you sit with mouse and keyboard you expect to be a little more in control. The ads our Consumer-ombudsman wants to ban are those that effectively take over over the browser - in the case of full screen ads, the whole box (After reading The Reg for maybe a couple of years I have got the feeling that you do not like people taking control of your machines either - be that Micro$oft, RIAA/MPPA, US govt, crackers or whoever else).

Sure, you can get rid of those by only enabling Java/Flash when you absolutely need it (banking, overblown websites) and killing ActiveX completely. You can also redirect annoying sites to localhost, or get some ad-blocking soft... I do all of the above, except the soft, but the proverbial less technically savvy only has that as an option that he does not know that he has, let alone how to choose. Leaving such people behind in flashy ad-dust as sheep for the sheering is a kind of tech-darwinism that won't do this world any good in my opinion.

Please don't post my email-address. I don't like ads popping up in my inbox either.

Oscar Røhling

For the record, here's why we don't have pop-ups on The Register, a piece by Reg editor Drew Cullen. A great read 30 months on.

DNS in a Can?

Re: Feds investigating 'largest ever' Internet attack
and our earlier alert - Root server DoS attack slows net

why not dvd, 5 gigs of compressed zones is .. well pretty much all the root
zones:

ipsec# ls -la *.zone
-rw-r--r-- 1 root wheel 609901 Aug 23 16:14 edu.zone
-rw-r--r-- 1 root wheel 282403 Aug 23 16:14 gov.zone
-rw-r--r-- 1 root wheel 46241 Aug 23 16:14 inaddr.zone

ipsec# gzip -9 *.zone

ipsec# ls -la *.zone.gz
-rw-r--r-- 1 root wheel 164280 Aug 23 16:14 edu.zone.gz
-rw-r--r-- 1 root wheel 29486 Aug 23 16:14 gov.zone.gz
-rw-r--r-- 1 root wheel 1826 Aug 23 16:14 inaddr.zone.gz

Please note the .com/etc zones contain even more repetitive data (i.e. thousands of domains hosted via company X's 3 servers), and bzip will compress text even better then gzip).

In any event who cares, most major ISP's have pretty much all the important root zones locally anyways.

What worries me is things like BGP problems and attacks, a lot of weird things happen in BGP-land, is it simple network problems, or someone testing things? No BGP, no Internet (plus you can attack a lot more selectively with certain techniques and take out only certain providers, like say the US
military, or a country).

Kurt Seifried

Sandia

Re: 16,000 Opterons in Sandia supercomputer

Paul Komarek writes:-

As a numerical computing person, I think I'd make the same decision if I could verify claims about the
Opteron.

Here's why:

1) If you're doing massive multiprocessing, you need lots of cpus at the lowest price possible (and spend the money you save on a low-latency, high-bandwidth interconnect like those that Cray builds). That kills the Itanium.

2) The P4 has an extremely crappy floating point unit. I've written FP code for which a P4 Xeon 1.7GHz runs over *80* times slower than an Athlon XP 1900+. Many people have contacted me about this, audited the code, and repeated the experiment. The P4 fp unit simply stinks. If you use the sse2 unit on the P4 (sse1 is 32 bit, but sse2 is 64 bit) for this code, you'll see a >100 times speedup (making it faster than the XP 1900+ (as one would expect, considering the price of a P4 Xeon system versus an Athlon XP system). The downside is that the sse2 unit can only do the trivial stuff like multiply and add (I don't even know if it can do division, and if it can, how fast). So you still need the fp unit for other stuff. And using the sse2 unit means creating less-portable binaries

-Paul Komarek

Stay tuned - we keep missing each other.

---

A very important point missed in the Opteron vs Itanium debate is Performance/Watt. Itanium is a 120 Watt part. Now consider 16,000 of these. Let's see 16,000x120Watts = 1,920,000 Watts. LOL!

Opteron delivers more performance per Watt than Itanium, and I believe that was a major factor in addition to what you stated.

John R Bloch

Thanks for your letters. Some goodies coming up on software alternatives to Outlook inspired by Lotus Agenda, and unfortunately (for my ISP), those dangerous server pictures just keep on coming. ®