Has anyone considered the notion that Mono should be implemented because a Common Language Runtime is actually a good idea?
That's probably true, but approaches like PyPy and Parrot are probably going to be vastly superior in terms of both flexibility and performance (I'd bet on PyPy if I had to pick one right now--of course it only runs Python, Javascript, and a couple other languages at the moment, but it's got smart developers, serious European Union funding, and a design that lends itself easily to implementing other language front ends, and performance is improving at an incredible rate).
On the other hand, in the real world with real code, tools like BicycleRepairMan work great.
Of course, I could have code that evals a strong, or changes the base class of a live object, or alters the inheritance hierarchy, or whatever in my Python code. That's incredibly uncommon in good code, though, and it's up to me as a developer to know when I'm playing games like that and account for them.
In practice, for renaming a class/method/attribute, pulling up/pushing down methods and variables, etc BRM works find on 99% of my code--and the times that it doesn't are times that are obvious to anyone who knows anything about the system (it's not as though it silently fails in cases you'd reasonably expect it to work). And it integrates nicely with emacs and vim.
Don't forget that PyPy performance has improved by a couple of orders of magnitude over the past 6 months, and it seems likely that Python on PyPy will overtake C# for raw native performance within the year (PyPy has a large amount of European Union funding behind it, and the psyco Python JIT author is one of the core developers). Although for a lot of CPU-intensive tasks if you're doing numerics, image manipulation, etc then your python project is going to spend most of its time in hand-coded C libraries anyway.
PyPy is pretty darned cool; with the approach they took it should be a fast general virtual machine/JIT for many dynamically typed languages (they already have a JavaScript front-end as well).
Step 1: Make sure it compiles without errors. (This is a given if you were about to check it in). Step 2: Make sure that the name you're going to use is unique (use find and grep). Step 3: Go to the header where the symbol is defined and rename the symbol. Step 4: Compile with make -k; fix errors; repeat until it compiles without errors.
Good luck modifying code in a dynamically typed language that way (Python, Perl, Ruby, Smalltalk, etc). Just use a refactoring tool (a la BicycleRepairMan for Python).
The part of a machine that gets upgraded/replaced most often these days is the video card
I highly doubt this. The vast majority of machines out there aren't used for intensive gaming, and outside of that and a few graphics/engineering niches there hasn't been much of a reason to upgrade a video card in years.
I'm still using the same video card I've had for 6 years (largely because it has all the tv capture/out features I need and okay 3d performance but is completely fanless and silent).
An urban legend doesn't have to be false. From the alt.folklore.urban FAQ:
An urban legend does not have to be false, although most are. ULs often have a basis in fact, but it's their life after-the-fact (particularly in reference to the second and third points) that gives them particular interest.
See also the 'A Digression On Urban Legends And "Falsehood"' section of that FAQ (which also contains similar quotes from authors of published books on urban legends).
Over here you never know what's going to happen. One week the episode is on, the next it's replaced by something else, then you are watching a re-run from a previous season, then you are watching back-to-back episodes or back-to-back-to-back-to-back episodes. It's bizarre.
There's this moronical thing called "sweeps week" 4 times a year (February, May, July, November). Local advertisers set their rates for the rest of the year based on the networks' ratings in sweeps week. Consequently, networks have an interest in having great ratings that week, even if it means their overall ratings are somewhat lower (you're better off revenue-wise with, say, 120,000 viewers in sweeps week and 90,000 viewers the other 12 weeks of the quarter for an average of around 92,300 than you would be with 95,000 viewers every week).
One consequence is that networks will often have a popular show stop airing new episodes for a few weeks before sweeps week, and then they can promote the heck out of "Next week on an all-new Lost!" during sweeps week.
There's no good reason that local advertisers do their ratings this way instead of looking at week-to-week ratings. It's just tradition. National advertisers switched to week-to-week ratings long ago.
Family Guy, however, I never did understand why they cancelled it. It had a decent timeslot
???
I tried to watch Family Guy, but it was never in the same timeslot for more than 2 weeks at a time. It got pushed around every night of the week, and would be on for 2 weeks, then have a several week hiatus, then air another episode on a different night.
Futurama at least had a couple months of being on after the Simpsons before they started jerking it around. Family guy might've been in that spot once or twice but not with any regularity.
By "in the real world", I mean the average person off the street. From what I can see, they will react differently to a painting that they know is Van Gogh, or a movie that they know is a Spielberg, than they will to the same work if it's presented anonymously.
But there's a boatload of psychology that goes along with being "great art"
Don't you mean bullshitting? Many people lead shitty lives, but no one values the crap they produce just because they led a shitty life. If you have to inject a lot of backstory into a work of art to make it great.. well it must not be really that great to begin with.
I don't know what you mean by bullshitting, but it's the same stuff that causes people to value a family heirloom more than a brand new object that's exactly the same but not falling apart. Is that irrational? Yeah, for sure. And is that object "not really that great to begin with"? Sure. But I still value my grandmother's old radio flyer wagon more than the same wagon purchased new, or your grandmother's old wagon. Even if it is rationally just some old junk.
It's an internal value that people put on the perceived associations something has. It's why people will pay a bunch of money for a Beethoven manuscript, but not anything near the same amount for a nearly identical reproduction. It's why Topps can print a Mickey Mantle baseball card in 1952 and see it worth a small fortune today, but if they fired up the same presses and printed the same card today it wouldn't have the same value.
The sad thing about this is that people look to detect 'fakes' in order that they can be weeded out and derided as "not good". But if an artist can paint as well as one of the Masters, shouldn't we be excited to find a 'fake' because it means that there is another great painter out there who we know nothing about.
Not necessarily. A huge amount of art is tied up in personalities, and feelings of connection. It's a massively psychological field, not a rational one at all. A lot of the time, the artist's life, times, struggles, etc have as much to do with the impact of their art as the painting/sculpture/whatever does.
If you're just looking for something nice to look at, you'd be absolutely right. But there's a boatload of psychology that goes along with being "great art"--part of what people get out of looking at a Van Gogh is because they know his tragic story. Give them an _exact_ reproduction, but let them know that it's a reproduction, and the impact is less. Give them a so-so reproduction that they can't tell is fake--but that lacks a lot of artistry, and that experts would laugh at--and have them believe it's his, and the impact is much greater than the faithful reproduction.
And while it's irrational, I'm sure it's true and I'm not sure it's not valuable.
Pollock to me is a great example of the connection between the art and the actual artist having some bizarre sort of importance. To me, he's a legitimate artist--but his art isn't the paintings, it's his descriptions of the emotions and methods he uses to paint. The paintings themselves don't actually evoke any of those emotions, and they're basically pretty meaningless.
BUT, as an observer if I'm looking at something I think he painted, it helps me evoke those emotions because I "feel" connected. It doesn't actually matter whether he painted it or not--and looking at something he did actually paint that I've been told is a counterfeit is as bad as looking at a known fake.
It's purely psychological and completely irrational, but I really don't think I'm alone.
The name is often the only thing that matters to people.
Not just the name but what one could call (critically) artificial scarcity or (more supportively) some sense of connection with a time, place, or person. I mean, there's no reason that you couldn't make more baseball cards just like the 1952 Mickie Mantle rookie card. But even with the same name, printed by the same company, it's not "the real deal". You could say that a bubble-gum card has any real tie-in with the baseball player other than his photo is laughable, but another equally good (or even better, with modern technology) copy of the same photo lacks the value.
I think it's silly. But I also buy into it to an extent--I know that I have family heirlooms that are things which are still made, but having the original has value to me for no real reason. And I know when I go to the Carnegie Museum in Pittsburgh and look at Rodin's The Hand of God, it moves me--and if I looked at an identical reproduction, but knew it was a repro, I'd be less moved. Heck, even if I looked at the real thing but _thought_ it was a repro, I'd be less moved. That's surely irrational. Doesn't make it less true, though.
This season we had 3 of the 6 strongest hurricanes ever recorded (by pressure).
Kind of, but also kind of misleading because the numbers prior to 1970 are almost all pressure at landfall (rather than lowest pressure including off the coast). And if you count tropical cyclones worldwide then the strongest hurricane of 2005 (Wilma) is tied for 19th place.
By deaths, 1780 is the worst year with 3 storms killing over 1000 people--including one that killed 22,000 and caused such damage that late-arriving observers mistakenly believed it had been accompanied by an earthquake.
It's quite interesting that you say that whiskys are widely available in the US. (I'm from Germany, btw.) US friends of mine who live in rural areas have even problems to buy proper wines without problems.
Selection varies from area to area. Where I am, it's quite good; for instance, the Brickskeller pub downtown has over 700 kinds of bottled beer, including 70 or so from Germany. http://www.thebrickskeller.net/_wsn/page3.html has the complete list. That's certainly the widest selection for a pub in the area, but there are several other local pubs with large selection (including other places like Lost Dog with hundreds of bottled beers, and places like DA's RFD, Montgomery Grill, Bilbo Baggins, etc that have 10, 20, or 30 taps with interesting options from around the world).
If you're purchasing to take home, my local wine/beer store has 8000 different varieties of wine and over 1000 kinds of beer from all over the world.
And for liquor, I can find 30-40 different good single-malt Scotch whiskys at liquor stores within 15 minutes. One bar in the area has about 20 on their menu.
In rural areas, selection is a lot more limited. Still, I grew up in Maine and the local pub had 50+ beers on tap (lots of regional brews, but lots from around the world as well). If you ever go to Maine, check out http://www.greatlostbear.com/
Shipping wine between US states is illegal and getting good wines to one's home is often a real problem, so I thought strong alcohol is even harder to come by.
The US Supreme Court (in Michigan Beer and Wine Wholesalers Association v. Heald) struck down bans on interstate wine shipment in May of this year.
States can still ban wine shipment completely, but the earlier situation where they'd allow in-state shipment and ban interstate shipment was ruled illegal. About 20 states ban it completely, but many of the major states (California, New York, Texas, etc) allow wine/beer/liquor shipment--although there are often restrictions on how much you can get (2 cases/month per individual is common but it varies from state to state).
I prefer my Islays to be on the "smoke/heather" end of things rather than the "seaweed/iodine" end of things, so I prefer things like Bowmore and Lagavulin over things like Ardbeg and Laphroaig.
The difference of the casks is definitely cool, Glenmorangie (who currently own Arbeg as well) produce their eponymous Highland whisky in a number of different casks--sherry, port, burgundy, madeira, etc.
Read D.E. Knuth's "Structured Programming with Go To" (1974) for one paper showing how properly used goto can improve readability, maintainability, and performance.
Or read " 'Goto Considered Harmful' considered harmful"
In fact, performance is very much on their radar. One of the main developers is Armin Rigo, the author of psyco (specializing Python JIT that speeds up some kinds of python by a factor of 100 or so). Another is Christian Tismer, the author of Stackless Python and many pretty cool performance enhancements for the stackless Python interpreter.
Their approach was to get a correct implementation first, then begin work on optimization. In July or so they finally got an implementation that passes almost all of the Python test cases (and those that it fails on are actually testing implementation details of the CPython interpreter, not language features). At the time it was about 200x slower than the standard CPython interpreter.
However, the whole reason for the design they've picked is to allow for easy optimization. After July, the optimization effort began in earnest. Even without attacking the JIT/psyco-style stuff, they've already narrowed the gap to where it's now less than 7x slower than CPython. My guess is that within 6 months it'll be at least twice as fast as CPython for standard code.
The best part is that with the RPython approach they've taken, many of the speedups will apply not only to python but to other language front ends (e.g. there is an RPython implementation of a Javascript interpreter that should take advantage of all the optimization work they're doing).
3 more things: 1. You could easily get 12 people together and get 6 good bottles, and come out of it for under $20 a head to try 6 good options (or a hair more if you wanted one bottle of something ridiculous like Lagavulin or Macallan 25). 2. Another reason the Lagavulin lasts a while is that I generally prefer Talisker; Lagavulin requires the right mood (it's really smoky, while Talisker has almost a black-pepper finish instead--both are peaty, though). 3. Oban is mid-way between the Speysides and the Islay/Skye whiskys; I prefer going one way or the other, and Oban is pricey, but if you're looking for a transition and find a steal on it somewhere it could be worth keeping in mind.
Slashdot Mirror
Has anyone considered the notion that Mono should be implemented because a Common Language Runtime is actually a good idea?
That's probably true, but approaches like PyPy and Parrot are probably going to be vastly superior in terms of both flexibility and performance (I'd bet on PyPy if I had to pick one right now--of course it only runs Python, Javascript, and a couple other languages at the moment, but it's got smart developers, serious European Union funding, and a design that lends itself easily to implementing other language front ends, and performance is improving at an incredible rate).
On the other hand, in the real world with real code, tools like BicycleRepairMan work great.
Of course, I could have code that evals a strong, or changes the base class of a live object, or alters the inheritance hierarchy, or whatever in my Python code. That's incredibly uncommon in good code, though, and it's up to me as a developer to know when I'm playing games like that and account for them.
In practice, for renaming a class/method/attribute, pulling up/pushing down methods and variables, etc BRM works find on 99% of my code--and the times that it doesn't are times that are obvious to anyone who knows anything about the system (it's not as though it silently fails in cases you'd reasonably expect it to work). And it integrates nicely with emacs and vim.
Don't forget that PyPy performance has improved by a couple of orders of magnitude over the past 6 months, and it seems likely that Python on PyPy will overtake C# for raw native performance within the year (PyPy has a large amount of European Union funding behind it, and the psyco Python JIT author is one of the core developers). Although for a lot of CPU-intensive tasks if you're doing numerics, image manipulation, etc then your python project is going to spend most of its time in hand-coded C libraries anyway.
PyPy is pretty darned cool; with the approach they took it should be a fast general virtual machine/JIT for many dynamically typed languages (they already have a JavaScript front-end as well).
Step 1: Make sure it compiles without errors. (This is a given if you were about to check it in).
Step 2: Make sure that the name you're going to use is unique (use find and grep).
Step 3: Go to the header where the symbol is defined and rename the symbol.
Step 4: Compile with make -k; fix errors; repeat until it compiles without errors.
Good luck modifying code in a dynamically typed language that way (Python, Perl, Ruby, Smalltalk, etc). Just use a refactoring tool (a la BicycleRepairMan for Python).
The part of a machine that gets upgraded/replaced most often these days is the video card
I highly doubt this. The vast majority of machines out there aren't used for intensive gaming, and outside of that and a few graphics/engineering niches there hasn't been much of a reason to upgrade a video card in years.
I'm still using the same video card I've had for 6 years (largely because it has all the tv capture/out features I need and okay 3d performance but is completely fanless and silent).
An urban legend doesn't have to be false. From the alt.folklore.urban FAQ:
An urban legend does not have to be false, although most are. ULs often have a basis in fact, but it's their life after-the-fact (particularly in reference to the second and third points) that gives them particular interest.
See also the 'A Digression On Urban Legends And "Falsehood"' section of that FAQ (which also contains similar quotes from authors of published books on urban legends).
That's not a crime, it's a breach of contract (which is a civil and not criminal matter).
Over here you never know what's going to happen. One week the episode is on, the next it's replaced by something else, then you are watching a re-run from a previous season, then you are watching back-to-back episodes or back-to-back-to-back-to-back episodes. It's bizarre.
There's this moronical thing called "sweeps week" 4 times a year (February, May, July, November). Local advertisers set their rates for the rest of the year based on the networks' ratings in sweeps week. Consequently, networks have an interest in having great ratings that week, even if it means their overall ratings are somewhat lower (you're better off revenue-wise with, say, 120,000 viewers in sweeps week and 90,000 viewers the other 12 weeks of the quarter for an average of around 92,300 than you would be with 95,000 viewers every week).
One consequence is that networks will often have a popular show stop airing new episodes for a few weeks before sweeps week, and then they can promote the heck out of "Next week on an all-new Lost!" during sweeps week.
There's no good reason that local advertisers do their ratings this way instead of looking at week-to-week ratings. It's just tradition. National advertisers switched to week-to-week ratings long ago.
Family Guy, however, I never did understand why they cancelled it. It had a decent timeslot
???
I tried to watch Family Guy, but it was never in the same timeslot for more than 2 weeks at a time. It got pushed around every night of the week, and would be on for 2 weeks, then have a several week hiatus, then air another episode on a different night.
Futurama at least had a couple months of being on after the Simpsons before they started jerking it around. Family guy might've been in that spot once or twice but not with any regularity.
Battery life and reslilience. No moving parts, that's flash's one advantage that results in both benefits.
No moving parts also means relative silence and lower heat generation.
How do you know how accurate they are? Have you ever received your graded test back along with an answer key to compare? I've never seen that happen.
Yes, you can get your graded SAT back with an answer key to compare (I did it). It's a nominal fee.
By "in the real world", I mean the average person off the street. From what I can see, they will react differently to a painting that they know is Van Gogh, or a movie that they know is a Spielberg, than they will to the same work if it's presented anonymously.
The greatness of a work of art is not greater because of the suffering or backstory of the artist. Kill the artist, the value of art is in itself.
I disagree, strongly. In the real world, people's perception of the artist greatly alters how the art affects them.
I don't know what you mean by bullshitting, but it's the same stuff that causes people to value a family heirloom more than a brand new object that's exactly the same but not falling apart. Is that irrational? Yeah, for sure. And is that object "not really that great to begin with"? Sure. But I still value my grandmother's old radio flyer wagon more than the same wagon purchased new, or your grandmother's old wagon. Even if it is rationally just some old junk.
It's an internal value that people put on the perceived associations something has. It's why people will pay a bunch of money for a Beethoven manuscript, but not anything near the same amount for a nearly identical reproduction. It's why Topps can print a Mickey Mantle baseball card in 1952 and see it worth a small fortune today, but if they fired up the same presses and printed the same card today it wouldn't have the same value.
The sad thing about this is that people look to detect
'fakes' in order that they can be weeded out and derided
as "not good". But if an artist can paint as well as one
of the Masters, shouldn't we be excited to find a 'fake'
because it means that there is another great painter out
there who we know nothing about.
Not necessarily. A huge amount of art is tied up in personalities, and feelings of connection. It's a massively psychological field, not a rational one at all. A lot of the time, the artist's life, times, struggles, etc have as much to do with the impact of their art as the painting/sculpture/whatever does.
If you're just looking for something nice to look at, you'd be absolutely right. But there's a boatload of psychology that goes along with being "great art"--part of what people get out of looking at a Van Gogh is because they know his tragic story. Give them an _exact_ reproduction, but let them know that it's a reproduction, and the impact is less. Give them a so-so reproduction that they can't tell is fake--but that lacks a lot of artistry, and that experts would laugh at--and have them believe it's his, and the impact is much greater than the faithful reproduction.
And while it's irrational, I'm sure it's true and I'm not sure it's not valuable.
Pollock to me is a great example of the connection between the art and the actual artist having some bizarre sort of importance. To me, he's a legitimate artist--but his art isn't the paintings, it's his descriptions of the emotions and methods he uses to paint. The paintings themselves don't actually evoke any of those emotions, and they're basically pretty meaningless.
BUT, as an observer if I'm looking at something I think he painted, it helps me evoke those emotions because I "feel" connected. It doesn't actually matter whether he painted it or not--and looking at something he did actually paint that I've been told is a counterfeit is as bad as looking at a known fake.
It's purely psychological and completely irrational, but I really don't think I'm alone.
The name is often the only thing that matters to people.
Not just the name but what one could call (critically) artificial scarcity or (more supportively) some sense of connection with a time, place, or person. I mean, there's no reason that you couldn't make more baseball cards just like the 1952 Mickie Mantle rookie card. But even with the same name, printed by the same company, it's not "the real deal". You could say that a bubble-gum card has any real tie-in with the baseball player other than his photo is laughable, but another equally good (or even better, with modern technology) copy of the same photo lacks the value.
I think it's silly. But I also buy into it to an extent--I know that I have family heirlooms that are things which are still made, but having the original has value to me for no real reason. And I know when I go to the Carnegie Museum in Pittsburgh and look at Rodin's The Hand of God, it moves me--and if I looked at an identical reproduction, but knew it was a repro, I'd be less moved. Heck, even if I looked at the real thing but _thought_ it was a repro, I'd be less moved. That's surely irrational. Doesn't make it less true, though.
This season we had 3 of the 6 strongest hurricanes ever recorded (by pressure).
Kind of, but also kind of misleading because the numbers prior to 1970 are almost all pressure at landfall (rather than lowest pressure including off the coast). And if you count tropical cyclones worldwide then the strongest hurricane of 2005 (Wilma) is tied for 19th place.
By deaths, 1780 is the worst year with 3 storms killing over 1000 people--including one that killed 22,000 and caused such damage that late-arriving observers mistakenly believed it had been accompanied by an earthquake.
Schlitz tastes like hair.
It's quite interesting that you say that whiskys are widely available in the US. (I'm from Germany, btw.) US friends of mine who live in rural areas have even problems to buy proper wines without problems.
Selection varies from area to area. Where I am, it's quite good; for instance, the Brickskeller pub downtown has over 700 kinds of bottled beer, including 70 or so from Germany. http://www.thebrickskeller.net/_wsn/page3.html has the complete list. That's certainly the widest selection for a pub in the area, but there are several other local pubs with large selection (including other places like Lost Dog with hundreds of bottled beers, and places like DA's RFD, Montgomery Grill, Bilbo Baggins, etc that have 10, 20, or 30 taps with interesting options from around the world).
If you're purchasing to take home, my local wine/beer store has 8000 different varieties of wine and over 1000 kinds of beer from all over the world.
And for liquor, I can find 30-40 different good single-malt Scotch whiskys at liquor stores within 15 minutes. One bar in the area has about 20 on their menu.
In rural areas, selection is a lot more limited. Still, I grew up in Maine and the local pub had 50+ beers on tap (lots of regional brews, but lots from around the world as well). If you ever go to Maine, check out http://www.greatlostbear.com/
Shipping wine between US states is illegal and getting good wines to one's home is often a real problem, so I thought strong alcohol is even harder to come by.
The US Supreme Court (in Michigan Beer and Wine Wholesalers Association v. Heald) struck down bans on interstate wine shipment in May of this year.
States can still ban wine shipment completely, but the earlier situation where they'd allow in-state shipment and ban interstate shipment was ruled illegal. About 20 states ban it completely, but many of the major states (California, New York, Texas, etc) allow wine/beer/liquor shipment--although there are often restrictions on how much you can get (2 cases/month per individual is common but it varies from state to state).
Do you mean Ardbeg? I've never heard of "ardbeck". Ardbeg is an Islay that's pretty widely available in the US.
e s/Ardbeg.JPG
This one:
http://www.maison-de-stuff.net/john/whisky/pictur
I prefer my Islays to be on the "smoke/heather" end of things rather than the "seaweed/iodine" end of things, so I prefer things like Bowmore and Lagavulin over things like Ardbeg and Laphroaig.
The difference of the casks is definitely cool, Glenmorangie (who currently own Arbeg as well) produce their eponymous Highland whisky in a number of different casks--sherry, port, burgundy, madeira, etc.
What? The language with a *goto* construct wins?
Read D.E. Knuth's "Structured Programming with Go To" (1974) for one paper showing how properly used goto can improve readability, maintainability, and performance.
Or read " 'Goto Considered Harmful' considered harmful"
In fact, performance is very much on their radar. One of the main developers is Armin Rigo, the author of psyco (specializing Python JIT that speeds up some kinds of python by a factor of 100 or so). Another is Christian Tismer, the author of Stackless Python and many pretty cool performance enhancements for the stackless Python interpreter.
l og.20051213.
Their approach was to get a correct implementation first, then begin work on optimization. In July or so they finally got an implementation that passes almost all of the Python test cases (and those that it fails on are actually testing implementation details of the CPython interpreter, not language features). At the time it was about 200x slower than the standard CPython interpreter.
However, the whole reason for the design they've picked is to allow for easy optimization. After July, the optimization effort began in earnest. Even without attacking the JIT/psyco-style stuff, they've already narrowed the gap to where it's now less than 7x slower than CPython. My guess is that within 6 months it'll be at least twice as fast as CPython for standard code.
Armin has an interesting overview of how the JIT stuff will fit in in the IRC logs at http://tismerysoft.de/pypy/irc-logs/pypy/%23pypy.
The best part is that with the RPython approach they've taken, many of the speedups will apply not only to python but to other language front ends (e.g. there is an RPython implementation of a Javascript interpreter that should take advantage of all the optimization work they're doing).
Your boss is an idiot if he thinks more than one person can work on code without a revision control system
Heck, I think I'm a pretty good coder but I wouldn't work on anything non-trivial by myself without some sort of versioning.
3 more things:
1. You could easily get 12 people together and get 6 good bottles, and come out of it for under $20 a head to try 6 good options (or a hair more if you wanted one bottle of something ridiculous like Lagavulin or Macallan 25).
2. Another reason the Lagavulin lasts a while is that I generally prefer Talisker; Lagavulin requires the right mood (it's really smoky, while Talisker has almost a black-pepper finish instead--both are peaty, though).
3. Oban is mid-way between the Speysides and the Islay/Skye whiskys; I prefer going one way or the other, and Oban is pricey, but if you're looking for a transition and find a steal on it somewhere it could be worth keeping in mind.