Even Windows modal dialogs could be moved (a modal dialog is one that pops up and demands attention by not letting you interact with the rest of the system - often used for error notifications and similar).
What you're describing is the art of uselessness, like old splash screens that obscure the things you're working with - but at least they fade after a few seconds.
I have the same problem in many respects, I have a lot of C++ skills, yet I was turned down for a position writing C++ on Linux because I didn't have enough Linux experience - admittedly, I have little commercial development experience on Linux (there's not many jobs for Linux over the last 15 years) but I have a fair amount of sysadmin exp on the platform.
But I guess, the little boxes weren't checked and so I failed to get past the recruiter gateway.
which matters hugely if all you consider from your programmers are how cheap they are and how much you can sell them on for (and if their code is crap and it costs extra to maintain the product - w00t, that just makes you even more money).
and I'm sure you get a lot of hassle from the kids who come to you to ask how various things are done.
Its the same everywhere I've worked, there's always a group of older workers who are the go-to guys if you need to now how something works, or if you need advice on how to put your stuff in the bigger picture.
The biggest problem for me is the crap the kids come up with - for example, I recently was shown a new web service that had 1 method on it, which was implemented using 6 interfaces and 10 files. And this had a comment saying "I didn't use dependency injection because this is such a simple project". It was the hallmark of someone who's taken on every OO way of working with factories and wrappers and decided to use them all without the experience to know when to use them.
the reason for this is Windows - because these programs need to roll their own updaters, they tend to write the code to be as cross-platform as the program is. So there's no incentive for them to scrap code they know and replace it with a different updater.
It might make sense for the end-user, and I'd hope the manufacturers of these programs would start to support more default package managers now they're getting more into Linux, but they're still really Windows programs for the most part.
don't worry - no doubt they'll ditch an old, old version of OpenOffice for the latest version of Microsoft Office.. you know, the one that has a new interface that doesn't allow you to scroll your documents vertically, making you swipe them sideways like you would on a tablet.
Its the old rock and a hard place problem, only they could upgrade to the current LibreOffice version. Ho well, their loss.
what happened to the write-once, run anywhere concept?
See, if Google made Android as part OS features, part Java API, then you could run an updated app against an old phone and the new features you expect to be present would simply not work - it'd throw an exception at runtime if the user attempted to use the missing API call (assuming the dev didn't look for and hide that option).
Android fragmentation isn't any more of a problem than the existing problem of having lots of phones running different stock Android versions.
Windows 8 can be seen a a lot of things, and I doubt Sinofsky was the sole dictator who had everything made to his personal design. that's not quite the way things work, however....
guess who the new head of Windows div is? Julie "Ribbon" Larson-Green..... also the same person who was responsible for... well, guess.
"Unlike other companies that maybe have one person at the top, we don't have a [design] czar at Microsoft," says Julie Larson-Green, VP of program management for Windows. Of Metro, she adds, "Its not like Steve [Ballmer] decreed it." One former longtime Microsoft manager put it bluntly: "I don't think Steve could even spell the word design." And unlike Steve Jobs, who was infamous for meddling in every detail of Apple's product launches, Ballmer didn't go to any of the rehearsals at Milk Studios for the unveiling of the Surface; his part was played by a stand-in till he arrived on the day.
So if the brass were so indifferent to design, how did this thinking emerge at Microsoft at all?
In May 2009, Julie Larson-Green corralled 150 thought leaders from various Microsoft groups (Office, Phone, Bing, Xbox) in the Redmond, Washington, campus conference center to kick off planning for Windows 8.
you want to see TFS 2012... its one large blank canvas with a few bits of text here and there, no rhyme or reason for it, just minimalist and unhelpful. It also doesn't help that the menus are deliberately hidden away with no real hints that there's something to click on (on the web view that is). In the explorer, the menu items are scattered all over various menus - try setting security on the iterations.. hint: its not in the "security settings" menu item.
Microsoft will take this opportunity to focus on the cloud as a replacement for everything, driven by the 30% fee app store.
Come to think of it - the floating blocks are a start for this - if your product has an arbitrary set of disconnected features, then its easy for something else to add to them.
A lot of people say that - Sinofsky is seen as "being difficult to work with", but given that at various places where he gave tech presentations and was ready to answer any question, you have to think "difficult" means "an effective leader who actually does stuff" (besides brown-nose superiors and empire-build that is) and that pissed off the political arse-lickers.
as a client, I use Mint (cinnamon) at the moment. For servers I always put Centos on and run it headless - ie command line only.
Centos is the free clone of RedHat so its very popular in corporates. You might like Fedora for a client if you use it, as that's RedHat's "bleeding edge" distro where they test out the new stuff before its fully stable, when it gets put into the server version, RHEL.
or, it was something to do with the successes he had.... which reminds me, why did Bob Muglia leave... was it:
a) because he was useless, only taking Server and Tools from a cost to a billion-dollar sales engine? b) Because of his communication style? c) because of his inability to plan for the future? d) because he was a shoe-in as Ballmer's replacement when the shareholders kick him out.
have you seen VS2012? I mean I use VS2010 every day and I just had to install the SQL server 2008r2 components to make it view sql files as text!!!
That's a bit bloated, it may appear fast but if you look at how much RAM it's using, and all the components its loading you'll see its just as bloated as Eclipse... which is admittedly Java so I agree is bloated crap. Eclipse works just as fast as VS nowadays, and I know, I run both - though I do run eclipse as a PHP and C++ IDE, not Java.
The point though, they're just as bad as each other nowadays. The days of fast and slim VC6 are well gone.
Windows Server already is command-line only, but you can still install the GUI as an option..... which every Windows admin will.
the thing to bing for (haha, only kidding - google for) is Windows Server Core. Here's a beginners guide to it, so you can now install Windows Server on underpowered hardware, like that old dual-core i3 you were thinking of chucking out:)
One thing to note that i found really interesting is this : There is no support for managed code. so don't expect it to be a GUI-less version of Windows, its much more a cut-down thing for specialist uses, or for basic OS features like file+print.
For example one of my neighbors is a well regarded author... she's still using office 2003 and will not switch to a newer version because she just can't tolerate the tool ribbon and she says most authors whom she knows feel the same way
this is my problem with Office and the whole Microsoft monoculture... not that she refuses to upgrade - frankly, progress happens, UIs change, whatever. No, my problem is that she's using Word at all as if there were no other word processors in existence.
For an author, I'd recommend Scrivener. Its a word processor type program but with a lot more features geared toward managing a very large document, whereas Word is really a pretty poor system for writing shitty business reports.
understand that you're coming from a position of ignorance and frankly, it shows!
Windows may not be as good as the alternatives in many respects, but its not totally useless. It has a very capable shell - Powershell, you can install all manner of IDEs without the monstrosity that is visual studio (yeah, it has bloated a lot), but there's code:blocks, eclipse, qt's suite, intel's compiler even. It can display folders with preview pictures - in many different styles, like Large icons, medium or small icons, or tiles etc.
So... if you're going to give it a go, you have to give it a proper chance. Your post is like a Windows user installing Linux and complaining there's no way to map network drive letters.
yeah, but this guy is a manager at Cisco.. the skills he acquirted over his very long career involve licking his superior's ass, stuffing his face with dinner and charging to expenses, stabbing his colleagues in the back and generally not having a clue about anything other than his own ambition.
so in this case its more "I don't know who you are. I don't know what you want. if you are looking for ransom,. I have lots and lots of money but I won't give any of it away. But what I do have are a very particular set of skills; skills I have acquired over a very long career. Skills that make me a laughing stock for people like you. If you let my pride go now, that'll be the end of it. I will still look for you, I will still try to sack you. But if you don't, I will look for you, I will find you - probably behind the photocopier which is where I found my set of postit notes - and I will tell you off, so there".
I totally agree with that sentiment. If the kernel boys aren't going to offer some binary compatibility between versions then Nvidia should just write a wrapper for the kernel that calls into the proprietary driver internals.
ah, but you're not talking of TDD as found in many systems, you're talking about TDD as it was originally envisaged.... they're well different. So much so that the term BDD is almost what TDD used to be before the auto-generation tool took over and made everyone think they were doing test-driven development by clicking the "make tests from my code" button that creates a set of stubs, 1 per method.
[sarcasm]Isn't reflection great.. just look at the cool things you do with it... [/sarcasm]
so anyway, I dislike TDD because of this, though it has its place as a way to easily put little "checking" tests in. For the serious stuff, we use a couple of "BDD" tools like dbfit, behat and cucumber.
most post-apocalyptic civilisations will have a form of power source based on an internal combustion engine, though they're likely to be called Ermintrude or Daisy and you'd best not be anywhere the exhaust when the internals are combusting.
Maybe they will build this tractor in the shape of 2 large pieces of metal roped together with a wooden (or other renewable-tech) handle to provide a means of guidance.
does it come with an easy-to-install gasoline manufacturing plant too?
When I was a kid we did stuff for Sierra Leone, and one part I remember was the bit where well-meaning western organisations would fund raise to buy tractors for the place, only that a year later these would be rusting away after various mechanical failures in the heat and dust, or because the villager couldn't afford to run them. What the villagers really needed, and no-one in the industrialised world figured out, was shovels.
The problem is often that you look for an answer to your own perceptions, and no those of the end-user.
there are fewer programmers around than there used to be, maybe because kids today all want to do media studies rather than anything difficult, but it means the pool of talent is smaller. There is also an aspect where the kid who can make more noise about how great product xyz is can appear to be an expert, and persuading management that he is competent isn't something they'd know any different. when faced with a team lead who has no real knowledge either, it can be a killer combination - for incompetence that still looks good. Today, its easy to say "yes but performance always was going to be poor given the complexity of the uber difficult business logic we're using, we'll just have to scale out into the cloud", and you've hit 5 buzzwords which makes you a winner! collect your promotion to architect at the door:)
I do like using lots of tools that are good at what they do, rather than "its integrated so we'll use it".. meaning "we'll give Microsoft even more money for a half-arsed attempt at a product that has way better alternatives". Currently I'm using TFS... jenkins+SVN/git+Redmine would kick its sorry arse, but... hey, it's integrated in VS so it must be better. pah.
Glad to give you a nice quote though, please use it freely:)
I'm not convinced by TDD though, I do it but I've never really needed it in all the years of coding I've been doing. I prefer to write up test harnesses that exercise more of the system in a larger granularity. ie, instead of writing tests that exercise a method, I prefer to exercise a class - and then the test I write can be an example of how to use that class if it involves setting up, configuring it to whatever task you want, and then making it do work.
I also find this helps find more bugs than traditional TDD, eg the time I had a network class that had methods to set ip address and port, but if you set the port first, it would fail (as setting ip would first initialise the entire internal address variables). TDD doesn't find those bugs, and they're the ones I'm more interested in.
Scrum is really bad... I wish we were doing scrum,. instead we're doing DSDM (yes, you'll have to look it up). Its worse than scrum as it tries to fit a defined timescale to agile processes. So instead of prioritising and working until you've done as much as you can get done, you have to continually worry about whether you can get all you've agreed to done in time. Still, our stand-ups are too long, and our stories are either too small or too large, and some of the team don't want to apply tasks to them. So we don't have accurate burndowns either!
They'll get it eventually though, or management will enforce it to the book (and the expensive consultancy that was brought in to teach everyone this crap.....)
Its what it is though, I'll stick with it until I really get fed up, you can't just dump something without trying to change it.
I don't think there was - he chose an engne to work on the platform everyone makes games for - Windows. then he adds the poll option "d'ya want it on Linux?" and 21% of responders have said "yes".
That does mean 79% think other factors are more important.
so whilst I think it would make sense to think linux at the start of every project, especially now that Steam's opinion is known, it wasn't the position he started from. I wonder how much work has already started with CryEngine, and whether its too late to change, or too late to get CryEngine updated to work with Linux proper (if it already works on Android, I can't see it being too difficult a task).
No, we do as a team get together to plan, but we do far too much of it That means that planning, not doing, is what we're spending all our time doing. Our agile processes are bolloxed up.
the concept of disappearing for 2 months to solve a problem is actually pure agile. It may not work for you, and it may not work for your organisation in which case it's not the process you want (agile doesn't mean - follow this exact process or else).
Alistair Cockburn (one of the original agile movement people) has this to say about agile:
One member in the Crystal family of methodologies is Crystal Clear. Crystal Clear can be described to a Level 3 listener in the following words:
"Put 4-6 people in a room with workstations and whiteboards and access to the users. Have them deliver running, tested software to the users every one or two months, and otherwise leave them alone."
That is agile. Really. The fact that agile can be other things doesn't detract from it. sure, you might not be able to do that, or you might not want to do that.. but it is still agile.
One thing that is definitely not agile is getting bogged down in processes. Agile is all about freeing yourself from the processes that prevent you from delivering what you're supposed to be working on. Planning, documentation (unless it's required by the business need), specifications, test plans, meetings, all of these things are overhead - you might need them, you might need to do them to some extent. But you must recognise that they are not what you're supposed to be doing and as such should be minimised.
Too often the planning and meetings take precedence over delivery. That's exactly what agile was developed to overcome. Take a look at the agile manifesto.
Slashdot Mirror
Even Windows modal dialogs could be moved (a modal dialog is one that pops up and demands attention by not letting you interact with the rest of the system - often used for error notifications and similar).
What you're describing is the art of uselessness, like old splash screens that obscure the things you're working with - but at least they fade after a few seconds.
The java... its a good thing :)
I have the same problem in many respects, I have a lot of C++ skills, yet I was turned down for a position writing C++ on Linux because I didn't have enough Linux experience - admittedly, I have little commercial development experience on Linux (there's not many jobs for Linux over the last 15 years) but I have a fair amount of sysadmin exp on the platform.
But I guess, the little boxes weren't checked and so I failed to get past the recruiter gateway.
which matters hugely if all you consider from your programmers are how cheap they are and how much you can sell them on for (and if their code is crap and it costs extra to maintain the product - w00t, that just makes you even more money).
and I'm sure you get a lot of hassle from the kids who come to you to ask how various things are done.
Its the same everywhere I've worked, there's always a group of older workers who are the go-to guys if you need to now how something works, or if you need advice on how to put your stuff in the bigger picture.
The biggest problem for me is the crap the kids come up with - for example, I recently was shown a new web service that had 1 method on it, which was implemented using 6 interfaces and 10 files. And this had a comment saying "I didn't use dependency injection because this is such a simple project". It was the hallmark of someone who's taken on every OO way of working with factories and wrappers and decided to use them all without the experience to know when to use them.
the reason for this is Windows - because these programs need to roll their own updaters, they tend to write the code to be as cross-platform as the program is. So there's no incentive for them to scrap code they know and replace it with a different updater.
It might make sense for the end-user, and I'd hope the manufacturers of these programs would start to support more default package managers now they're getting more into Linux, but they're still really Windows programs for the most part.
don't worry - no doubt they'll ditch an old, old version of OpenOffice for the latest version of Microsoft Office.. you know, the one that has a new interface that doesn't allow you to scroll your documents vertically, making you swipe them sideways like you would on a tablet.
Its the old rock and a hard place problem, only they could upgrade to the current LibreOffice version. Ho well, their loss.
what happened to the write-once, run anywhere concept?
See, if Google made Android as part OS features, part Java API, then you could run an updated app against an old phone and the new features you expect to be present would simply not work - it'd throw an exception at runtime if the user attempted to use the missing API call (assuming the dev didn't look for and hide that option).
Android fragmentation isn't any more of a problem than the existing problem of having lots of phones running different stock Android versions.
Windows 8 can be seen a a lot of things, and I doubt Sinofsky was the sole dictator who had everything made to his personal design. that's not quite the way things work, however....
guess who the new head of Windows div is? Julie "Ribbon" Larson-Green..... also the same person who was responsible for... well, guess.
"Unlike other companies that maybe have one person at the top, we don't have a [design] czar at Microsoft," says Julie Larson-Green, VP of program management for Windows. Of Metro, she adds, "Its not like Steve [Ballmer] decreed it." One former longtime Microsoft manager put it bluntly: "I don't think Steve could even spell the word design." And unlike Steve Jobs, who was infamous for meddling in every detail of Apple's product launches, Ballmer didn't go to any of the rehearsals at Milk Studios for the unveiling of the Surface; his part was played by a stand-in till he arrived on the day.
So if the brass were so indifferent to design, how did this thinking emerge at Microsoft at all?
In May 2009, Julie Larson-Green corralled 150 thought leaders from various Microsoft groups (Office, Phone, Bing, Xbox) in the Redmond, Washington, campus conference center to kick off planning for Windows 8.
you want to see TFS 2012... its one large blank canvas with a few bits of text here and there, no rhyme or reason for it, just minimalist and unhelpful. It also doesn't help that the menus are deliberately hidden away with no real hints that there's something to click on (on the web view that is). In the explorer, the menu items are scattered all over various menus - try setting security on the iterations.. hint: its not in the "security settings" menu item.
Microsoft will take this opportunity to focus on the cloud as a replacement for everything, driven by the 30% fee app store.
Come to think of it - the floating blocks are a start for this - if your product has an arbitrary set of disconnected features, then its easy for something else to add to them.
A lot of people say that - Sinofsky is seen as "being difficult to work with", but given that at various places where he gave tech presentations and was ready to answer any question, you have to think "difficult" means "an effective leader who actually does stuff" (besides brown-nose superiors and empire-build that is) and that pissed off the political arse-lickers.
as a client, I use Mint (cinnamon) at the moment. For servers I always put Centos on and run it headless - ie command line only.
Centos is the free clone of RedHat so its very popular in corporates. You might like Fedora for a client if you use it, as that's RedHat's "bleeding edge" distro where they test out the new stuff before its fully stable, when it gets put into the server version, RHEL.
or, it was something to do with the successes he had.... which reminds me, why did Bob Muglia leave... was it:
a) because he was useless, only taking Server and Tools from a cost to a billion-dollar sales engine?
b) Because of his communication style?
c) because of his inability to plan for the future?
d) because he was a shoe-in as Ballmer's replacement when the shareholders kick him out.
have you seen VS2012? I mean I use VS2010 every day and I just had to install the SQL server 2008r2 components to make it view sql files as text!!!
That's a bit bloated, it may appear fast but if you look at how much RAM it's using, and all the components its loading you'll see its just as bloated as Eclipse... which is admittedly Java so I agree is bloated crap. Eclipse works just as fast as VS nowadays, and I know, I run both - though I do run eclipse as a PHP and C++ IDE, not Java.
The point though, they're just as bad as each other nowadays. The days of fast and slim VC6 are well gone.
Windows Server already is command-line only, but you can still install the GUI as an option..... which every Windows admin will.
the thing to bing for (haha, only kidding - google for) is Windows Server Core. Here's a beginners guide to it, so you can now install Windows Server on underpowered hardware, like that old dual-core i3 you were thinking of chucking out :)
One thing to note that i found really interesting is this : There is no support for managed code. so don't expect it to be a GUI-less version of Windows, its much more a cut-down thing for specialist uses, or for basic OS features like file+print.
For example one of my neighbors is a well regarded author ... she's still using office 2003 and will not switch to a newer version because she just can't tolerate the tool ribbon and she says most authors whom she knows feel the same way
this is my problem with Office and the whole Microsoft monoculture... not that she refuses to upgrade - frankly, progress happens, UIs change, whatever. No, my problem is that she's using Word at all as if there were no other word processors in existence.
For an author, I'd recommend Scrivener. Its a word processor type program but with a lot more features geared toward managing a very large document, whereas Word is really a pretty poor system for writing shitty business reports.
understand that you're coming from a position of ignorance and frankly, it shows!
Windows may not be as good as the alternatives in many respects, but its not totally useless. It has a very capable shell - Powershell, you can install all manner of IDEs without the monstrosity that is visual studio (yeah, it has bloated a lot), but there's code:blocks, eclipse, qt's suite, intel's compiler even. It can display folders with preview pictures - in many different styles, like Large icons, medium or small icons, or tiles etc.
So... if you're going to give it a go, you have to give it a proper chance. Your post is like a Windows user installing Linux and complaining there's no way to map network drive letters.
yeah, but this guy is a manager at Cisco.. the skills he acquirted over his very long career involve licking his superior's ass, stuffing his face with dinner and charging to expenses, stabbing his colleagues in the back and generally not having a clue about anything other than his own ambition.
so in this case its more "I don't know who you are. I don't know what you want. if you are looking for ransom,. I have lots and lots of money but I won't give any of it away. But what I do have are a very particular set of skills; skills I have acquired over a very long career. Skills that make me a laughing stock for people like you. If you let my pride go now, that'll be the end of it. I will still look for you, I will still try to sack you. But if you don't, I will look for you, I will find you - probably behind the photocopier which is where I found my set of postit notes - and I will tell you off, so there".
I totally agree with that sentiment. If the kernel boys aren't going to offer some binary compatibility between versions then Nvidia should just write a wrapper for the kernel that calls into the proprietary driver internals.
Like ndiswrapper but for graphics.
ah, but you're not talking of TDD as found in many systems, you're talking about TDD as it was originally envisaged.... they're well different. So much so that the term BDD is almost what TDD used to be before the auto-generation tool took over and made everyone think they were doing test-driven development by clicking the "make tests from my code" button that creates a set of stubs, 1 per method.
[sarcasm]Isn't reflection great.. just look at the cool things you do with it... [/sarcasm]
so anyway, I dislike TDD because of this, though it has its place as a way to easily put little "checking" tests in. For the serious stuff, we use a couple of "BDD" tools like dbfit, behat and cucumber.
most post-apocalyptic civilisations will have a form of power source based on an internal combustion engine, though they're likely to be called Ermintrude or Daisy and you'd best not be anywhere the exhaust when the internals are combusting.
Maybe they will build this tractor in the shape of 2 large pieces of metal roped together with a wooden (or other renewable-tech) handle to provide a means of guidance.
But I guess such things aren't "cool" enough.
does it come with an easy-to-install gasoline manufacturing plant too?
When I was a kid we did stuff for Sierra Leone, and one part I remember was the bit where well-meaning western organisations would fund raise to buy tractors for the place, only that a year later these would be rusting away after various mechanical failures in the heat and dust, or because the villager couldn't afford to run them. What the villagers really needed, and no-one in the industrialised world figured out, was shovels.
The problem is often that you look for an answer to your own perceptions, and no those of the end-user.
there are fewer programmers around than there used to be, maybe because kids today all want to do media studies rather than anything difficult, but it means the pool of talent is smaller. There is also an aspect where the kid who can make more noise about how great product xyz is can appear to be an expert, and persuading management that he is competent isn't something they'd know any different. when faced with a team lead who has no real knowledge either, it can be a killer combination - for incompetence that still looks good. Today, its easy to say "yes but performance always was going to be poor given the complexity of the uber difficult business logic we're using, we'll just have to scale out into the cloud", and you've hit 5 buzzwords which makes you a winner! collect your promotion to architect at the door :)
I do like using lots of tools that are good at what they do, rather than "its integrated so we'll use it".. meaning "we'll give Microsoft even more money for a half-arsed attempt at a product that has way better alternatives". Currently I'm using TFS... jenkins+SVN/git+Redmine would kick its sorry arse, but ... hey, it's integrated in VS so it must be better. pah.
Glad to give you a nice quote though, please use it freely :)
I'm not convinced by TDD though, I do it but I've never really needed it in all the years of coding I've been doing. I prefer to write up test harnesses that exercise more of the system in a larger granularity. ie, instead of writing tests that exercise a method, I prefer to exercise a class - and then the test I write can be an example of how to use that class if it involves setting up, configuring it to whatever task you want, and then making it do work.
I also find this helps find more bugs than traditional TDD, eg the time I had a network class that had methods to set ip address and port, but if you set the port first, it would fail (as setting ip would first initialise the entire internal address variables). TDD doesn't find those bugs, and they're the ones I'm more interested in.
Scrum is really bad... I wish we were doing scrum,. instead we're doing DSDM (yes, you'll have to look it up). Its worse than scrum as it tries to fit a defined timescale to agile processes. So instead of prioritising and working until you've done as much as you can get done, you have to continually worry about whether you can get all you've agreed to done in time. Still, our stand-ups are too long, and our stories are either too small or too large, and some of the team don't want to apply tasks to them. So we don't have accurate burndowns either!
They'll get it eventually though, or management will enforce it to the book (and the expensive consultancy that was brought in to teach everyone this crap.....)
Its what it is though, I'll stick with it until I really get fed up, you can't just dump something without trying to change it.
I don't think there was - he chose an engne to work on the platform everyone makes games for - Windows. then he adds the poll option "d'ya want it on Linux?" and 21% of responders have said "yes".
That does mean 79% think other factors are more important.
so whilst I think it would make sense to think linux at the start of every project, especially now that Steam's opinion is known, it wasn't the position he started from. I wonder how much work has already started with CryEngine, and whether its too late to change, or too late to get CryEngine updated to work with Linux proper (if it already works on Android, I can't see it being too difficult a task).
No, we do as a team get together to plan, but we do far too much of it That means that planning, not doing, is what we're spending all our time doing. Our agile processes are bolloxed up.
the concept of disappearing for 2 months to solve a problem is actually pure agile. It may not work for you, and it may not work for your organisation in which case it's not the process you want (agile doesn't mean - follow this exact process or else).
Alistair Cockburn (one of the original agile movement people) has this to say about agile:
One member in the Crystal family of methodologies is Crystal Clear. Crystal Clear can be described to a Level 3 listener in the following words:
"Put 4-6 people in a room with workstations and whiteboards and access to the users. Have them deliver running, tested software to the users every one or two months, and otherwise leave them alone."
That is agile. Really. The fact that agile can be other things doesn't detract from it. sure, you might not be able to do that, or you might not want to do that.. but it is still agile.
One thing that is definitely not agile is getting bogged down in processes. Agile is all about freeing yourself from the processes that prevent you from delivering what you're supposed to be working on. Planning, documentation (unless it's required by the business need), specifications, test plans, meetings, all of these things are overhead - you might need them, you might need to do them to some extent. But you must recognise that they are not what you're supposed to be doing and as such should be minimised.
Too often the planning and meetings take precedence over delivery. That's exactly what agile was developed to overcome. Take a look at the agile manifesto.