Saturday, March 31, 2012

Top 3 books read in 2011

Yes, it's rather late for this sort of thing: it's the last day of March and therefore the final day of the first quarter of 2012.  Somehow drawing attention to the fact that it's still the first quarter of the new year makes it seem like this isn't really as late as it is.

I didn't pay that much attention last year to the books I read, so it's all going to be a bit vague, but here, briefly, are the three best books I read last year:

In order of reading.

The Quantum Thief, by Hannu Rajaniemi (2010)
This one stands out a little for me, selection wise, in that it was a very recent publication.  Though I have a bit of a bent these days for modern SF, that normally means reading something from the last couple of decades.
There was a fair bit of hype on this debut and it doesn't fail to deliver.  Hannu writes a very dense, very hard post singularity sci-fi novel that throws some brilliant concepts around some pretty cool characters.  Post-humans abound, working with even more powerful entities, in and around the solar system perhaps a few centuries hence.  Contains detectives, thieves, gaming culture and a whole lot of musings on the implications of future tech.
Looking forward to the sequel coming out later this year, but will need a reread of the Quantum Thief beforehand, methinks.
Trivia: contains a recommendation on the cover by Charles Stross, who wrote the very next book I read and is next up on this list.

Accelerando, by Charles Stross (2005)
Selected by me for a bookclub, since I'd been meaning to get around to Stross at some point and this one received some notoriety.  The book is a collection of 9 continuing tales, though it didn't suffer for that.  Taken together, they form a novel focusing on a man at the cutting edge of technology, and his descendents, as they head into and then through the technological Singularity, beginning in roughly the present day.  Slashdot makes good background reading for this (and even gets a mention), especially for the first third that deals with the ramp up to the Singularity.
Probably has the highest density of ideas to pages of any book I can recall reading.  Stross throws down concepts and references without even bothering to stop and see if the reader is still with him because he's already bringing something else up.

Pushing Ice, by Alastair Reynolds (2005)
I'd picked up a slew of books by Reynolds at a garage sale recently and had rather expected that the first Reynolds books that I'd read would be the Revelation Space series.  The first book in that series was in fact sitting in my short-short list of books awaiting to be read, whereas this book had languished in a more general pool of books: the second tier, as it were.  Still, for whatever reason this book was picked up in the dying weeks of the year and wham! it delivers emphatically with a strong plot and characters, along with a high degree of page-turning readability.  It has an adventure feel to it, and is quite reminiscent of Clarke's Rama series.  A crew of comet-miners are in the vicinity of an unknown alien object that's discovered and they are dispatched to investigate.  Cue humans interacting with cool alien tech.


General Notes

All three books are the only novels I've read by those authors, which also helps me tick off on another list a couple of the authors that I wanted to read something - anything - by.
All of the books are sci-fi, and in fact all are hard sci-fi.
All of the books are quite recent, being published within the previous 6 years.
Only one book (The Quantum Thief) is set in a specific time - the other two span non-trivial ranges.
All of the authors are European, and two of the three are from the UK.


Comments on the above novels are welcome, as well as your own favourite reads from last year, whatever the genres.

Thursday, March 8, 2012

Visual C++ 11 Beta 1 Initial Thoughts

I've started giving Visual C++ 11 a bit of a workout, and there's much to like about it.

Stability might be improved over 10 - I'll withhold judgement on that one for the time being, however.

We've got range based for loops.  That's a nice improvement since the DP last year.  Now you can do things such as :

#include

auto main()
-> int
{
    //  TODO - discover why this doesn't show up on Task List.
    std::vector joe;

    for (auto bob : joe)
    {

    }
}
In the above code we can simply declare bob as a deduced type of the container joe.  Handy when you're iterating over your containers and don't want to repeatedly deal with dereferencing (sorry, indirecting through, according to N3362) an iterator.

Pity there's such a mess with the Task List comments not working.  I can't get a single one to show.
And I won't even elaborate on the fact that it says it will only report on the ones in open files.  Sheesh: I've got a solution.  Feel free to inspect it.

The support for a light and dark theme is nice.  A little interesting that it's just a two horse event, but still.  Alas, I've kinda shot myself in the foot here by overriding some of the backgrounds in order to force a dark background - I'll have to undo some of that in order to get this fully working (should I want a light theme, and I might some day if I coded when the sun was up).

I really like the fully fleshed out syntax highlighting.  Though I think I loved it more in the Developer Preview.  I'm sure I used to be able to set things as italics - the example of parameters being in italics is mentioned across the web, however the Beta most definitely doesn't have an italics setting, just bold.  I have a function definition in front of me that's never known the loving hand of italicisation.

Oh, the monochrome look.  That's right.  For those who haven't heard, the Visual Studio now uses about three shades of grey as its entire colour palette.  Outside of your code, that really isn't an exaggeration.  There's a dark grey, a light grey and a - wait for it - intermediate grey.  Icons have been drained of colour and general boxing and lines have been removed from dialogs to give them a slabs-of-grey-paint look.
It was a little WTF at first.  You start thinking that this is a feature they'll introduce for VS 11 in order to have colour introduced as a feature in VS 12.  But I have to say that after the initial surprise, you get used to it pretty quick.  It's distinctive.  Better, worse?  Not much in it for me at this point.

Microsoft have folded a lot of the VS 2010 Productivity Power Tools into the VS 11 release.  The theory goes they get some out-of-band feedback on enhancements and iterate them, then add them into the next release - and here we are.  Alas, one of my favourites was the document view scrollbar.  And since I don't have the Power Tool available for VS 11, that means I'm without the enhanced scrollbar altogether.
I don't have a picture handy of the VS 2010 Power Tool one, so here's an image taken of the scrollbar from the Sublime Text editor home page.  Very much the same sort of thing: I love being able to see the general structure, as well as search matches, bookmarks, breakpoints, etc.  Sublime's not a bad text editor btw, though I tend to prefer notepad++.



 Overall, Visual Studio seems pretty punchy and responsive, though the find all references is as slow as ever.

Note that the Beta that you download is Ultimate.  Which in some ways is a pity because that means it comes with a boatload of stuff that I am never going to see in day to day usage (since I'd be using something like Professional).  Still, it does give you an opportunity to see what you're missing, including the perf tools and architecture stuff.

There are quite a few changes, both in the IDE and in the language support, so it will take quite some time to really go through and explore what's there.  This has just been a quick write up to gather some thoughts, prompted by the discovery of some unexpected C++11 support in the form of range based for loops.




Tuesday, February 14, 2012

On Latin and English and Mappings

It's illuminating how much of another language you can pick up if you've got a reasonable grasp of a related one.
And in the case of languages, there are a whole heap of related ones.
Now, I'm not talking about reading books, or even grasping full sentences.  But pithy quotes?  Not a total write-off.

The spark for this post was a Latin quote by a mate,
"Omnia mutantur, nihil interit"

But the origin of this post can be found some time ago.

When looking into the artificial language Ido, I came across a handy table describing words and the closest match in several languages.
One example was the noun, "kavalo", which in english means "horse".  So far, so confusing.  But the closest match shows that we can (imperfectly) map it to the concept of "cavalry", and this is indeed how I remember the word (along with the fact that nouns in Ido end in -o).

Fast forward to this evening and we have this latin phrase "Omnia mutantur, nihil interit".  Now, I'm not of an age where I learnt latin in school.  I'm also not of the bent that I decided to learn a dead language of my own accord.  I must admit I'm fond of the occasional latin phrase or word (readers will recall my hunt for ex-situ's antonym) but, all things told, I really don't know any latin worth mentioning.

The phrase was posted without an english translation, perhaps because it was assumed to be well known enough to not require it.  Alas, that was not the case for me.  And yet, looking over the phrase, mappings immediately became evident.
Omnia: omni- (prefix for all, everything omniscient, omnivore)
mutantur: mutate (to change)
nihil: nil (nothing)
interit: ... ( ...? )

So close, stymied at the last. 
Now, I'm vaguely familiar with the english translation, but even with the first three words I was unable to complete the phrase and had to resort to google.  "Everything changes, nothing is lost" for those of you still wondering.

So the question becomes, where is the link between the latin "interit" and the english "to lose, perish, decay"

The closest I can get is intermit, and that's not very close at all. 
Anyone?

Thursday, January 19, 2012

SOPA Oddity


SOPA Oddity

MPAA to User One
MPAA to User One
Take your videos and put your labels on
(Ten) MPAA (Nine) to User One (Eight)
Commencing upload, servers on
Check attribution and may God's love be with you.

This is User One under control, I'm looking at my screen
And I'm surfing in a most peculiar way
And the sites look very different today
Here am I sitting at a terminal, disconnected from the world
Lady Liberty is blue and there's nothing I can do

This is MPAA to User One, you really must obey
And the papers want to know whose files you share
Now it's time to post your comments if you dare

Though I'm past one hundred thousand views, I'm feeling very scared
And I think my blog just knows which links to post
Tell my ISP I love it much, it knows
Subscriber mail to User One, your server's dead, there's something wrong
Can you hear me, User One?
Can you hear me, User One?
Can you hear me, User One?
Can you...
Here am I sitting at a terminal, disconnected from the world
Lady Liberty is blue but there's something I can do...

Sunday, January 15, 2012

Unfinished Series

I'm not the sort of person who's entirely comfortable starting a series that hasn't been completed.  Many others seem to have no such issues, as demonstrated by the myriad people picking up books starting a new series within moments of them hitting shelves.  Myself, I usually only start an unfinished (non-open) series if the books are a gift or a loan.

I'm pretty sure it comes down to the fact that if I really enjoy it, I'm looking to see where it goes and to get some closure.  I don't have any problem with picking up books from an open ended series (Gor, Discworld) and reading them out of order.  I may have a slight preference for reading earlier works in the series before later ones, but not a strong compulsion.

At the moment I've got two unfinished series on the go, both unintentional.  Wait, that could be three if you count Rawn's The Ruins of Ambrai.  But given it's been 15 years since the second book I think we can safely discount that one.



  The first unfinished series I have going is the Penrose series by Tony Ballantyne.  I picked up the first book when it hit a mass market paperback on pre-order discount on the basis of a neat blurb and put it in the short list not realising it was part of a trilogy.  After reading several consecutive less than stellar books I needed something punchy, exciting and - above all - highly readable, so I grabbed Twisted Metal off the pile.  The book delivered the much needed invigoration to my reading, but towards the end I started to realise that the story wasn't really wrapping up.   Admittedly Ballantyne could have been taking a leaf out of Peter F Hamilton's book and leaving the start of the wrap-up for the final half a percent, but no, he's left it hanging for book two.
Penrose book two is already out (2010) and I've got a copy sitting in the short-short list, but I'm not seeing even a pre-order or title on the third book and the author described it late last year as "yet uncompleted".  Here's hoping it doesn't become a Captal's Tower.




 The second series I started accidentally was Hannu R's The Quantum Thief.  The book was getting some great press after its release and billed as Hard SF.  The novel itself is fantastic and flooded with killer concepts but again, upon reaching the end I found myself stung by the lack of closure.  Not too much, I have to admit, as a lot did wrap up, but it certainly had me scouring the net for more information on the trilogy.  Not as lucky here as I was with Ballantyne, as the second book is only due for release third quarter this year.

The question now is what to do about these series?
Well, specifically what to do about book two of Penrose.  I've got it lying around, and it's been maybe a year since I read the first book.  I thought maybe I'd read it about halfway between the first and the third book, but that's a little difficult to place, temporally speaking, when the third book hasn't been released.  Currently thinking about reading it soonish so the first one hasn't faded, with the expectation that the third comes out sometime next year so the second won't have faded by then.
The Fractal Prince, sequel to The Quantum Thief, I'll probably end up buying on release (paperback Octavo and HB will be published about the same time) and do a reread of the first book beforehand given the fact that you only start getting an idea of what the hell is going on when you're most of the way through.

And the above doesn't even touch on the issues of authors bringing out revised editions of their books (waves to Stephen King) or the fact that I'm actually considering reading A Song of Ice and Fire despite the series conclusion being perhaps a decade away...

Tuesday, December 20, 2011

Visual Complexity, by Manuel Lima

Coming to you from the "this is not a book review" department.




 Visual Complexity is probably one of the more surprisingly enjoyable books I've received recently.  And yet, it's rather quite unlike what I was expecting: and I say this as someone who had the book arrive in the mail with a large dose of "what the hell was this one?"  Perhaps a symptom of getting carried away with book ordering and having to wait for them to arrive from overseas, but nevertheless.



The book is something of an overview of different ways of trying to extract meaning from vast amounts of data through visual presentation.  The book shows how over time the forms employed have changed to suit the scale of the data being worked with.




One criticism I have seen is that given the amount of data on some of the graphs and the accompanying text, it is difficult to make out details of how the data is mapped.  I'm not overly bothered by it myself, as it's not the specifics of any particular graph that I'm interested in, more the style used for presentation. 

 

It's not going to teach you how to go about organising or analysing your data, but if you do have to deal with such things this book will show you the results of the myriad ways others have tackled similar problems.
All in all, recommended if you're looking for inspiration or pretty graphics.

Friday, December 2, 2011

Visual C++ 11 and Boost, Pt 2.

The modifications of the build.bat file described in the previous post give us better VC 11 compatibility, but there are still a few shortcomings that we need to deal with, particularly if we want to get auto-linking working.  With the changes described so far, Visual Studio will attempt to link to libraries from the previous version of visual studio (2010), resulting in complaints about missing libraries.

The remaining VC 11 issues are spread over two files.
The changes, as before, are mostly trivial and it's mostly a case of keeping track of those few places that need VC 11 cases.  No doubt a later boost version will have more significant changes to these files to more accurately track the changed capabilities of the compiler and standard library.

Compiler Capabilities (visualc.hpp)

First config/compiler/visualc.hpp

There are a number of issues with the file, but none serious and none dealing with auto-linking issues.
The purpose of the file is to specify for boost which C++ features are available or not available, given the current version of VC.  Of course, the version checking tops out at 2010.

We probably want to remove the warning that we're using an unknown compiler, a check performed at the very end.  We just need to bump the version up to 1700, as below:

// last known and checked version is now 1700 (VC 11):
#if (_MSC_VER > 1700)
#  if defined(BOOST_ASSERT_CONFIG)
#     error "Unknown compiler version - please run the configure tests and report the results"
#  else
#     pragma message("Unknown compiler version - please run the configure tests and report the results")
#  endif
#endif


Shortly above that the compiler version is defined, again topping out at VC2010.  We just need to modify that section to include a VC 11 case:

#   elif _MSC_VER == 1600
#     define BOOST_COMPILER_VERSION 10.0
#   elif _MSC_VER == 1700
#     define BOOST_COMPILER_VERSION 11.0


Finally, above this we have a set of defines for all versions of the compiler, specifying those features that are not supported in any VC release.  Doesn't look like this really changes for VC 11, but they're under the following banner in case you're curious:

// C++0x features not supported by any versions

Autolinking (auto_link.hpp)

We need boost to define the library toolset for us, and config/auto_link.hpp tops out at VC2010.  Once again a simple extra case for VC 11 (and modification of the previous case to stop it matching) is all we need.  Check that BOOST_MSVC is 1700 or greater for VC 11 before moving on to CBuilder 6 and auto link will once again try searching for the correct libs when required.  Trivial, as below:

#elif defined(BOOST_MSVC) && (BOOST_MSVC == 1600)
   // vc10:
#  define BOOST_LIB_TOOLSET "vc100"
#elif defined(BOOST_MSVC) && (BOOST_MSVC >= 1700)
   // vc11:
#  define BOOST_LIB_TOOLSET "vc110"
#elif defined(__BORLANDC__)

Conclusion

That pretty much concludes all the changes you should need to make to the boost build system to get it playing nice with VC 11.

Latest version of Boost (1.48)
http://www.boost.org/