Nobody is questioning the product, which as John has stated many times, is brilliant when working.
What the report is questioning is Canon USA's policies towards its customers when the printer isn't working.
I'm not even questioning the reliability of the printer, which I believe to be excellent. If a printer isn't working, once the printer is fixed, it appears to just keep on printing without a hitch. We aren't getting any reports of parts breaking over and over again. All of the reports are generally regarding defects that are present fairly near the time of purchase (although there is some legitimate concern that roll feed units could have the gears pop off at an inopportune time, like 366 days after you bought the printer).
The IPF5000 Wiki is not the center of the IPF5000 universe (although it may be for a small group of posters ). We have about 65 actual members and 300-400 unique visitors per day. That is a lot of lurkers. We don't know if they own the printers or are just looking. There are about 15 reports of defective roll feed units. That would be a 25% figure based on the membership (probably way too high, but who knows).
Also, I have some anecdotal information to suggest that the roll feed unit problem is fairly widespread:
1) A poster to the Wiki reported that 3 roll feed units failed for someone doing a demo of the IPF5000. Don't know over what time period (hours, days, weeks). The audience was apparently not impressed. The reporter is deemed to be a reliable source based on his posting history.
2) Another poster reported that the service tech who came to fix his roll feed unit said he was fixing 4 roll feed units every week. Again, the poster was deemed reliable.
All we can do on the Wiki is identify trends. Since there are many people who don't know about the Wiki, we really can't come up with any kind of denominator. We can only identify the trouble spots, but can't tell how big they are. Conversely, we can identify many areas that AREN'T trouble spots, e.g., there is not a single report of clogging to the Wiki, EVER.
For some reason people are drawing the conclusion that I don't like the printer. Nothing could be further from the truth. Now that I know its quirks and my printer is working properly I am producing beautiful prints. It has a number of really good points, which are listed in the article. However, the Wiki recommendation is based on the weight of the available evidence. With the information at hand, would I go out and purchase this printer right now? No. Would I recommend that one of my friends purchase it? No. I know there is a good chance the roll feed unit will have to be repaired, and that may require more than one service call (I know that sounds nuts, since Canon knows the problem is common and has redesigned the parts, but it has been reported by several people). Why would I recommend a printer known to have these problems (warranty, timely service, poor documentation, questions about longevity) unless it was so much better than the competition in other areas that the risk and hassle was worth it?
The one wild card in this is the question of print longevity. Since I don't sell my work, it isn't that big an issue to me, but for those who do I would be concerned. It has been 13 months since the initial estimate of "at least 100 years". Data on other printers has come and gone. I would be worried if this was important to me. I think this is actually the biggest cloud hanging over the printer. Past experience has shown that there can be large differences between ink sets, and big differences between particular ink and paper combinations. Note that HP has outstanding Wilhelm results, better than Epsons. It all depends how important that is to your needs.
Per a personal communication from Joseph Holmes (great landscape photographer and creator of Ekta Space
http://josephholmes.com ), the Wilhelm methodology does not account for "reciprocity failure". In other words, reducing light intensity does not produce an exactly corresponding increase in time. The factor to correct for this varies, but is thought to average around 2.5. Thus 100 "Wilhelm years" is probably more like 40 real years. For that reason, you may want a lot more "Wilhelm years" on your prints, depending on your application. Again, this is according to Joseph Holmes, and beyond the realm of my own knowledge.
One thing I am very pleased about is that the article is sparking a healthy dialog. My findings are not set in stone, and as I said in the article, if Canon makes changes to the problem areas you will see the Wiki recommendation change in a heartbeat. The reason I wrote and published the article was not to keep people from buying the printer or trash Canon. It was to try to spur Canon to action so I can again recommend purchase of the IPF5000!
--John