Comment on
By d.w.

There may be hope -- CBS bought the wonderful Last.fm last year, and they haven't ruined it yet.

Comment on DVR Lifestyle
By Ryan Wells

I don't think I'm recording anything from NBC right now, so I can't check if my DVR was affected.  I know this, though:  If I had to go back to watching TV without DVR time-shifting, I would simply stop watching all together.  Between iTunes and Netflix, I could get the shows I'm really interested it, and I'd probably save time and money besides.

Hmm...on second thought, that might actually be good for me, so maybe I should hope for more network idiocy! :)

Comment on Stupid
By Vincent Clement

I wouldn't be able to watch some of those shows if it weren't for Vista Media Center. If I am unable to watch these shows at my convenience, I'll just give up watching NBC shows. Idiots.

Comment on Broadcast flag also affects TiVo users
By Randal L. Schwartz

I haven't checked if any of my TiVo programs I taped monday were flagged (not even sure if I tape anything from NBC, for that matter).  But the broadcast flag definitely tags some of my recordings, and at that point, I can't use TiVo-to-go to grab the show to take it on the run.  Ugh.

Comment on most likely the drive went to sleep mode
By diego scataglini

Most likely than not one of this cases happened:

a) the drive or the mac went in sleep/stand-by mode and didn't unmount properly. 

b) The drive was unplugged without unmounting the drive. 

This happens a lot with linux operating systems. When the computer then tries to access it, the drive is visible but it appears as locked or in use by somebody else or otherwise inoperable.

Windows is much better at this although not perfect. When this happens I just plug the drive on a windows box. It would be recognized just fine and then I unplug it or do a safe hardware removal. After that the drive is ready to be connected again to a mac or a linux box.

Comment on most likely the drive went to sleep mode
By diego scataglini

Most likely than not one of this cases happened:

a) the drive or the mac went in sleep/stand-by mode and didn't unmount properly. 

b) The drive was unplugged without unmounting the drive. 

This happens a lot with linux operating systems. When the computer then tries to access it, the drive is visible but it appears as locked or in use by somebody else or otherwise inoperable.

Windows is much better at this although not perfect. When this happens I just plug the drive on a windows box. It would be recognized just fine and then I unplug it or do a safe hardware removal. After that the drive is ready to be connected again to a mac or a linux box.

Comment on Firmware
By Claus

Could also be an issue with the firmware. Though it is more likely that the partition table went haywire.

Comment on That's definitely NOT what I wrote
By Ravi

I did not write, or even imply, the situation you have described.

Data is there forever, applications come and go. Different applications use the data in different ways. To bundle all the ways data can be used (for a given data structure) into one class is not necessarily the best way to do things. That is why we have the concept of mix-ins or modules, partly to let different applications do different  things with the same data, based on need. When you are bundling together operations and data, you are either bundling all the operations that can be performed on the data by different applications into one class, or you are only including operations for one (or a few) application(s). Neither approach seems aesthetically pleasing.

Separating the data, which will stay longer than the applications used to access it, from the operations required by the applications, seems logical to me.

For example, data that is used for transactional systems, needs different types of actions (CRUD, validation, integrity checks, etc.) While data used for Data Warehousing needs different operations. Would we keep all these operations in the same file/module/class? Or even in the same environment?

Also, the only feature of OO that can not be obtained in other imperative/declarative programming languages is inheritance. And experience has shown that composition is preferable to inheritance. Deep inheritance hierarchies make a system rigid and inflexible to change. Hence, even the unique feature of OO seems of limited use.

Comment on Re: I'm with Blaine
By James Robertson

Comment by James Robertson

So if I have a data structure that 5 functions operate on, having them in their own file magically separates them from the data?

Wow

If that's what you believe, then please - stop writing code now, for all our sakes :)

Comment on Paul Graham is right
By Ravi

I believe Paul Graham understands the implications of the mainstream OO philosophy bettar than many of its practitioners. And, surprisingly, so does Chris Date, of relational data management fame.

It is amusing to see those who are so enamoured of a "separation of concerns" philosophy (as in MVC, for example) are also very enthusiastic about bundling data and operations together into a "class", not realizing that they are contradicting themselves by espousing the exact opposite of a separation of concenrs idea.

Comment on I could be wrong...
By Piers Cawley

Maybe I'm wrong, but I got the impression that that was precisely Steve's point. I thought he'd made it pretty much explicitly in the body of the talk.

Comment on Where's the line?
By Daniel Berger

If I pick up a magazine and read a paragraph is it stealing? How about a page? How about half of it? Almost all of it? All of it?

Comment on
By Reinout Heeck

We haven't used bundles for several years now, I documented some of the reasons on Travis' blog. That post is a bit outdated now since reason 1) I mention there is being addressed in VW 7.6.

Comment on Questions to your Question
By Travis Griggs

http://www.cincomsmalltalk.com/userblogs/travis/blogView?showComments=true&printTitle=ABundleofQuestions&entry=3388098191

Comment on
By ppavel

I tend to use bundles extensively. I try to keep packages small and meaningful (minimal independent unit of functionality) and use bundles to organize them into larger blocks. 

Comment on Re: Store Questions
By Dennis Smith

Comment by Dennis Smith

We use bundles -- exactly two of them.

We have Framework/Tools bundle

A Business application bundle.

In general, those who work on one bundle do very little work in the other, so it keeps things separated. We update store based on one or the other bundle. It also means we can build an image with different versions of Framework vs Applications.

Comment on Does It Matter?
By W^L+

For this former Cingular user, their tie-in with AT&T is toxic. The phone looks great, especially once it goes 3G. But the only state where I had reliable AT&T service was New Jersey. No matter where I was, their support line would tell me that I was in a known dead area, but they planned to add more towers there soon.

Until VZ or Sprint offer the iPhone, I'm not buying one.

Comment on
By Boris

Apple said they are indeed sold out, http://www.cnn.com/2008/TECH/ptech/05/12/iphone.sold.out.ap/index.html

Comment on
By Boris Popov

We use bundles purely as means of organizing a tree of packages by meaning, not by isolated functionality blocks or anything fancy like that. For example, basic tree of categories would do just fine and we'd probably just have 2-3 packages in total: main codebase, web stuff, development+tests.

Comment on
By Carl Gundel

Right.  It also leaves out Messages (so, Objects + Messages), which Alan Kay goes on record many times as saying is the really crucial part of OO that people miss.

Comment on Unfortunately Graham's not alone
By Rick DeNatale

The quote is closely aligned with the definition of OO as it was coined by Alan Kay. Unfortunately that meaning got hijacked by the abstract-data type community, spearheaded by C++ which implemented classes and virtual functions as an improved way to do "data processing", and abetted by Peter Wegner's re-definition of Object Orientation as "Objects + Classes + Inheritance," only the first of which is essential in Alan's conception of OO.

Comment on Re: Some Bugs just linger
By Troy Brumley

Comment by Troy Brumley

Your post made me think of a few things:

• First that some software lasts and lasts, and the bug in question seems pretty minor, which is probably a contributing factor to its longevity.
• These bugs can get fossilized in software when people working on code higher levels in the stack code around them. How many programs broke when this particular bug was fixed? Probably not many based on my reading of the description, but some other long lasting bugs become features as users and other programmers conform to the bug's behavior.
• At least this wasn't declared a feature, like most of Microsoft Word's bugs ;)

Comment on Who are those People then?
By Claus

A small introduction - this applies to the engineer team pictures as well - would be wonderful - it is much nicer to be able to put names to faces or vice versa.

Comment on
By Carl Gundel

Very, very cool.  I hope they make that Seaside calendar widget available.  :-)

Comment on We should switch
By George

For things we know grow exponentially (e.g. hard drive capacity, clock speed), we should just switch to logarithmic measurements.

Your story would then be, "I can still recall buying a second 7.6 Logo-Byte drive for my first x86 pc... I currently have over 12.5 LB of storage in my office, and HP is talking about 15 LB storage."  The mind would boggle a little less (though, of course, it very much should boggle).

Similarly, my old Apple ][+ ran at 6.03 LHz, whereas one of the quad-cores at work runs at a staggering 9.56 LHz.

Or we could multiply all those numbers by 10 to make audio engineers more comfortable: 60.3 LHz, 95.6 LHz, 76 LB, 125 LB.... "My processor runs 1.1 dB faster than yours!"

Comment on definition of fair use
By Troy Brumley

I guess these nimrods define fair use as "anything we completely control". Personal use of my own purchased media is something I have a right to expect, and your DVD example is certainly "fair use" by any rational understanding of the term.