OT: Why it is better to pretend you know nothing about computers

[Man at B&Q]

[e-mail]...

If you are going to be pedantic, it was about opening, not saving "File locking stops you from opening it in two editors"

MBQ

[Mark]

Cos Microsoft has a deal with virtually all hardware vendors to ensure that every box has a copy of windows already installed.

[Daniel James]

I think you'll find that Aeriel is the spirit in The Tempest, not MSND, so the others should have been Prospero, Miranda, Caliban, etc..

Cheers, Daniel.

[Bernard Peek]

The complication with Word is that it's designed to handle documents that are too large to fit in memory. It uses temporary files with names based on the name of the open file. It's feasible to work around that but it's more work.

[Gaz]

And coldsores and genital herpes are very similar. I very rarely hear people tell me that they can feel a bout of genital herpes coming on.

[Daniel James]

Yes, what word does is complex. It seems to keep the original file open for reading and locked against (at least) writing, and to record changes to that file in one (or more?) temporary files until the changed file is saved and/or closed. This not only helps it to manage files larger than could be read into memory but provides support its autosave facility and also for multiple levels of undo command.

There are certainly good reasons why word doesn't just open the file and read it into memory in its entirety, as (say) notepad does.

Notepad is essentially just a wrapper application around a Windows Edit control -- it reads the whole file into the buffer of the edit control after which it can safely close the input file (leaving it available for other applications to read and/or write without any locking from Notepad). That behaviour does mean, though, that two separate instances of notepad may be used to open and modify the file, and the changes from one will then be lost. Annoying (and confusing for the user who doesn't understand how it works) but a price worth paying for the simplicity of Notepad -- it's not intended to be sophisticated.

I'm not sure how Wordpad works ... experience suggests that it does not load the whole file into memory in one go -- it loads large files more quickly than Notepad, for example, and IIRC can load larger files than Notepad can (though maybe the file is still being loaded into memory; Notepad is limited in the size of file it can load by the size that its Edit control's buffer has been set to, which certainly isn't all of available memory).

.. to work around what? Word's use of temporary files? Why try to work around what is actually a pretty reasonable approach, especially given that the files are used in document recovery after a crash (which wouldn't be possible from any purely in-memory working space).

Cheers, Daniel.

[Mike Tomlinson]

In article , Jules Richardson writes

Oh, I'm sure there wasn't, or they'd still be in court arguing the toss now. Dave Cutler, one of the architects of NT, had a VAX/VMS background and probably took some of the concepts with him to M$.

I have also noticed that Acorn's RISC OS, for those who know of it, has a distinctly VAXish flavoured command line.

[Mike Tomlinson]

In article , Bob Eager writes

aaaaaaghh!

[Mike Tomlinson]

In article , Steve Firth writes

You haven't seen NT on Alpha, then. :o)

[GAP]

I beg to differ - when I was running it, NT on Alpha was a screamer, but that was because *at the time* my 64-bit Alpha was considerably faster than any 16-bit Intel box that was available. However the complications of getting the untranslated applications to run via the emulator (I've forgotten what it was called now) rather negated that performance boost... Still, I stuck with it for a year or two :-)

Graham (At that time a digital employee)

[GAP]

It would be a serious understatement to say that Dave Cutler (or God as he was considered inside digital :-)) had a VMS backgound!

My understanding was that the clustering algorithms did "make their way" into a certain west coast software company, but a licensing deal was done to retrospectively legalise their use. Sadly that company still hasn't produced a clustering product that works properly, as defined by the way it works in VMS.

Graham

[Andy Burns]

I tend to think 32 bit windows was available at that time!

FX!32

[Daniel James]

Yes, indeed, NT on x86 predates NT on Alpha.

However the point is that NT on Alpha was faster than NT on intel at the time, and I do recall seeing a very nice demo in which the same application (a complex-looking 3D modelling and rendering thing) was run side-by-side on a top-end Compaq x86 box and two dec Alpha boxes, one running the x86 code under emulation and one running native Alpha code. The point was made that all three boxes had the same graphics hardware, so as to make the test as fair as possible.

The Alpha emulating x86 was respectably faster than the x86 code running natively, while the Alpha code running natively ran rings around both of them.

Of course, it was dec that had set this demo up, so they will have picked an example that showed their fancy hardware in the best light .. but it was still impressive.

Cheers, Daniel.

[Mike Tomlinson]

In article , Daniel James writes

The Alphas were superb machines. We ran our network on them (using Tru64 UNIX, OSF/1 as-was) until about four years ago, and still use an AlphaServer DS10 running OpenVMS and a custom application for telescope pointing calculations.

The striking thing about DEC equipment was the build quality, just head and shoulders above anything else. Some of that ethos has fed through into HPaq's Proliant servers.

[Johny B Good]

Regarding the latter comparison, it would be more accurate to claim winXP as being the Idiot Son of win2k.