Python has some language features that one might never use, but then
one day you come across them in some other code and need to know what
they do. The following fragments (see the reply by Xoanan on this page) are
To determine the intersection between two lists "list1" and "list2":
intersection = filter(lambda x:x in list1, list2)
To determine the union between two lists "list1" and "list2":
union = list1 + filter(lambda x:x not in list1, list2)
To determine the difference between two lists "list1" and "list2":
difference = filter(lambda x:x in list2, list1)
To determine the distinct elements, those not in common between two
lists "list1" and "list2":
distinct = filter(lambda x:x in not list2, list1) + filter(lambda x:x in not list1, list2)
A discussion of the various ways of extracting the
unique elements from a list
From time to time you might encounter a list which contains some lists,
and you want to flatten this into a single list of simple elements.
artical talks about two ways to do this. Perhaps the more readible
method is with the nested list comprehensions:
nested = [[1,2,3], [4,5], ]
flatList = [x for sub in nested for x in sub]
[1, 2, 3, 4, 5, 6]
this works by the first "for" loop (for
sub in nested) iterating over the top list and on each iteration
picking up a sublist and placing a reference to it in "sub", then the
second for loop (for x in sub)
runs and picks up the selected sub list and iterates over it, placing
each element of it in x. A list of all the individual values that x
takes is built up by the [x ...]
construct. A limitation with this is that all the elements in the
outer list must support iteration (i.e. be lists, tupples or something
else list-like), so you cannot have a simple scalar element in the
A review of the Canon
9900 large format printer. As for the cost of ink here is some info
Speaking with Canon,
I found the
Of course, every
image is different, but in Canon's testing, the company used the ISO
Standard #5 image to estimate inktank usage. (This image provides a
fair and accurate ratio of colors to approximate the average digital
photograph.) Here are the results, according to Canon:
Cyan: 1100 pages
Yellow: 540 pages
Photo Cyan: 380
Photo Magenta: 280 pages
Green: 2300 pages
new red and
green tanks, obviously, are used far less often than the others because
they're primarily used for accent work. They're the last tanks you'd
have to replace.
Jason Bovberg -June 3, 2004
Based on CDN$20 per cartridge (current pricing seems to be in the $15 -
$19 range) this would work out as $0.24 per print. What the poster did
not mention was what the print size was... in a follow up it was said
to be 8x10 on standard printer settings, using Canon Photo Paper Pro
Here is some more information:
I purchased the i9900
about 6 weeks ago
to replace my S9000. The
S9000 is no slouch and a top rated printer from 2 years ago but i9900
is much better. I now shoot with a Canon 10D and Canon L lenses even
though I have thousands of transparencies to scan on a Canon FS4000US
film scanner. The 10D was purchased just prior to a trip to Rome where
I shot about 1000 images.
I just finished
assembling 170 of
the best Rome images into an album using Canon PhotoRecord software
that comes with most Canon digital products. I printed the album on
both Epson dual-sided matte paper and Pictorico dual-sided semi-gloss
paper. The results are amazing. Both papers work extremely well with
the i9900 although they each have their own subtle tinting difference.
This can be corrected easily with color adjustment through the printer
driver. Printing the entire album on the dualsided 8.5 x 11 stock (68
pages total) used no more than half of some of the ink cartridges and
almost none of the red and green. People who have viewed the album are
stunned and amazed with the quality and can't believe it came from an
inkjet printer. The color punch and tone is incredible and the level of
detail and resolution equals or exceeds wet chemistry printing. If you
want to see dots you'll need at least a 4X loupe and there is
absolutely no banding.
It is a mistake to
not consider this
printer because of the limited selection of Canon papers. Epson papers
work extremely well on the Canon printers as do Pictorico, Mitsubishi
and Konica. These are the only papers I've had the opportunity to try.
Needless to say, I am very happy
I bought this printer.
Bob Baron -June 18, 2004
Oct'01: AMD has released multi-processor versions of their new
compares it to the original 1200MHz version. And is there any real difference
between the Athlon and Athlon-MP chips?
An article that explains
the difference between a circular and a linear polarizing filter,
and goes into the potential colour shifts that can result. More on
polarizers as a filter to enhance infra-red photography.
of 16 USB flash drives with about 4GB of storage. This shows there
are still some large speed differences between the models. 
inkjet printers, the
890 and StylusPhoto
1280 (reviewed here)
which offer some of the best printing you can get. They can all do
borderless prints on regular media (no perforated tear-off strips needed), the
only real difference between the 780 and the 890 is that the 780 cannot
handle roll paper, the 1280 is the only one that can handle large format
paper. The I
Love Epson site has some information on new printers (probably in
Building a Babbage
Difference Engine out of a lot of Lego
 The PackardBell EasyNote XS makes an appearance at CES'08. From the pictures it appears to have an external DVI port. This appears to be the same as the Cloudbook from Everex, which will be available from Walmart in the US starting 25-Jan-08. Looks like ASUS will need to do some revising of the Eee to stay ahead of the game, while not feature for feature identical they do target the same price point with some significant feature differences which is likely to split the market that ASUS has to itself at the moment. 
The GPL is a license, not a contract, this article
on Slashdot) explains the difference and why the common worry that
entwining GPL code into a proprietary system could cause loss of rights
for the proprietary work is not correct.
An article on the difference between QA
 There has been some talk that Intel might enter the Eee competition with something called the Netbook, here are some possible pictures of this. And some more views of it, where it's being called the Eco PC. It has appeared in Malaysia where it will be called the SmartBook and is made by FTEC and there is about a $40 difference in price between the 7 and 9 inch screens. It is also going to be made by CTL and will be called the 2go PC.  The ASUS Eee has some subtle differences in the keyboards between the white and black models - the black keyboards are reported to have more travel.  not all USB flash drives are the same (discussed here on Slashdot. If you are looking for a speedy drive, especially if you need to write a lot of small files, then there can be large differences in performance and probably the only way t ofind a fast one is through tests. Kingston has some documentation about what sort of NAND flash RAM technology it uses in its products, which may be significant if you use a flash drive for something that does a lot of writes (like running an operating system off one). They also mention that flash drives can store data for up to 10 years under normal conditions - something that does not receive much attention. This is particularly troubling as prices for SD cards (the most common format for digital cameras) have dropped to the point (in mid-2008, some are going on sale for $10 for a 2GB card) that one could consider just using them on a shoot once and then file in an album basis and not bothering to transfer the contents to hard disk or DVD media.  If you thought that different HiDef channels (or shows) looked different you're not imagining things, apparently there are large differences in image qualities between HD stations these days.  Some SATA RAID controller card solutions.
Tom's Hardware takes a look at a number of SAS and SATA RAID controller cards.
The HighPoint RocketRAID 3520 SATA RAID controller card.
The Areca ARC-1231ML controller, this review compares it to the Promise SuperTrack STEX6850 and the ICH9R (Intel chipset solution that is often found on motherboards) controllers. Unfortunately they only examined RAID-0 and RAID-1 performance so did not find much in the way of differences. A comparison of nine Serial ATA RAID 5 adapters dates from 2005 but goes into a lot of details, including looking at differences between the CPU-hosted and on-board processor approaches.
 Build a web server on a business card sized PCB (discussed here on Slashdot). This uses a micro-SD card to hold the files that the web server will serve up. Sparkfun Electronics has a similar, pre-built, Mini-Web PIC Development Board for about $40, the main difference is that it only has 128Kbytes of on-board storage for the web pages - though if you were using this to implement parts of a home control system that would not be an issue. Another possibility are the EZ Web Lynx devices.  In Why RAID 5 Stops Working in 2009 the author argues that with the arrival of 2TB drives RAID arrays will have got large enough that the unrecoverable read error rate will put a RAID rebuild at significant risk. While this seems possible it would appear to be a rather poor software design if a RAID system completely fails a rebuild if a single read error happens, I would have thought that the logical thing to have done would have been to pause the rebuild and ask the operator if he wanted to stop or continue and accept the consequences (which would probably be a corrupted file). A comment on this issue that addresses some of the concerns. An article that looks at the causes of some of the drive errors. How the UNRAID system handles this issue gets discussed here, with UNRAID the data is not striped so a failure of this kind will not affect the whole array and, according to the developers, an error of this kind during an array rebuild will not stop the rebuild (though a file may be corrupted as a result). CERN took some time to research the silent data corruption problem. Slashdot has another discussion of the issue of RAID arrays becoming more likely to fully fail during a rebuild. Adaptec goes through some of the reliability calculations for RAID arrays here and here. A graduate student takes a look at RAID system reliability through simulations. Another article on the unrecoverable read error problem. There is a research project looking at read errors in Microsoft, an interim progress report is here. This article examines the probability of encountering a read error while rebuilding a RAID array and explains the math fairly well. The paper: An Analysis of Data Corruption in the Storage Stack takes a look at all sorts of error sources, incuding the UREs. Are there really differences between the SATA and more expensive SCSI and Fibre Channel drives? The importance of disk scrubbing as a way of keeping the UREs at bay. A white paper from Hitachi on the URE issue and why RAID6 makes sense (though it looks like their math might be off). Wikipedia's article on S.M.A.R.T. explains a lot of the SMART counters. Hard Disk MTBF: Flap or Farse calls into question the reliability of MTBF ratings, as does this paper.  An ad-hoc review of 21 different USB flash drives (from 1GB to 64GB). This got extended in a second part which looks at the speed difference between FAT32 and NTFS for these drives. Generally FAT32 was a bit faster than NTFS. Note that for transferring files between different computers, especially when they are in different domains, FAT32 is much easier to use.  An "after 6 months" report on switching to solar power for a house, this shows the difference between summer and winter electric generation. Some of the comments talk about the apparent error in this installation - the panels are mounted flat to the roof which does not allow for tracking within the day or across the seasons (which will affect power generation greatly). An update on how this is going after a year.  The limitations of computer math, though in some systems (like that of the Patriot missile control system) it is more limited than it should be. The cited error in time calculations of 0.34 seconds (which lead to an error in position of almost 700m) strikes me as a bit odd given that the time was being stored with a 0.1 second resolution. Normally when one is storing times in a computer system those times are obtained from a higher precision hardware timer circuit which might work to milli-second or even micro-second accuracy and only when the time is sampled by the computer is it rounded down to the 0.1 second resolution. If this is done, then each time stamp is accurate to +/-0.05 seconds and the difference between any two such sampled times has a maximum absolute error of 0.1 seconds (but a typical error of less than that). Now given a 0.1 second error the position should really have been accurate to within about 200m - the next question would be is that accurate enough?  Slashdot discusses the price difference between lithium ion batteries intended for laptops and power tools.  Some articles on the Weibull distribution which can be used to model the probability of failure of devices. The significant difference between this and the normal distribution is that the Weibull is a one-sided curve, so all the events happen above a reference point rather than being distributed on either side of the reference point. This makes sense for things like time to failure given that a component cannot fail before it is put into use (well, if you include manufacturing defects and shipping issues it could...).
  Mobilicity is a new cell phone provider that started service in May'10 in Toronto Canada. Their plans are all "unlimited" the only difference between them being the features carried. Unfortunately this means to get data from them one has to get all the other features too, meaning for any data you have to pay $65/month, even if you don't do much voice. 
['difference'] is in these pages: