Boot time

I think that’s more because of fragmentation, so it needs to jump to next free space without overwriting other files.

If Rick is talking about Windows transfer and he’s using the GUI then the problem is the swirly paper bit.
Trying a vbscript transfer will show a much faster transfer. All under the bonnet work rather than something I christened as “dancing dinosaur syndrome” back in the 90’s. Yes, fragmentation on windows systems has an impact on that type of thing but not to a great extent.

I’ve never seen a USB drive that cant write 3GB in an hour.

The Philips flash key managed to have glitches in a BBC SD broadcast. Given it’s about 3GB for a two hour film, in inclined to suspect the Flash stick is just crappy. I have a whizzy fast SD card, I’ll try that later.

Shame, I’ve just discovered that when the receiver dumps the transport stream to file, they mean that literally. I can switch on subtitles! ;-)

I think that’s more because of fragmentation, so it needs to jump to next free space without overwriting other files.

Shouldn’t be an issue for flash memory, no mechanics to move…

Shouldn’t be an issue for flash memory, no mechanics to move…

I don’t know much about flash, so take this with pinch of salt, but it emulates something like a disk, so the OS is convinced it needs to jump and those routines might be slow. In reality, nothing it where it seems in flash, and your flash chip has a little microcontroller crunching numbers trying to figure out where it should store the data without accidentally destroying nearby wafers.

so the OS is convinced it needs to jump and those routines might be slow

I rather suspect that any slowness of the routines would be vastly outweighed by the time it takes for a physical head seek. Drives use figures like 20ms seek… If your processor is running at a gigahertz, imagine how many instructions that 20ms would equal.
On the other hand, flash memory has delays in erasing blocks, which could be what is up with my USB key – a lacklustre flash chip that erases slowly…

If your processor is running at a gigahertz, imagine how many instructions that 20ms would equal.

but now we live in a generation were use of ultra-high level languages are okay, nobody codes a lot of what they need, they just pinch libraries and use only 5% of it.

Low-level programmers are getting rarer these days and they aren’t as valued so not only is it possible, it wouldn’t be surprising if there were millions of instruction in simple routines. I’ve seen more than my fair share.

Realistically though, I’d imagine any slow down is more to do with waiting for a resource or for another routine to return whatever and less likely that they is billions of instructions to process in one simple routine.

Low-level programmers are getting rarer these days and they aren’t as valued

There was a time when processors were sufficiently simple that human beings could hold in their heads all that was needed to write efficient assembly language code. With multiple level cacheing, out-of-order execution and other bells and whistles, I suspect that nowadays only the AI baked into compilers is capable of the job. Another step along the road to the redundancy of humans has been taken ;).

Another step along the road to the redundancy of humans has been taken ;).

Haha, but who knows how to write the compilers? Or are we doomed to use gcc, clang and vc++ long after developers die out and it’s no longer being updated?

Haha, but who knows how to write the compilers?

The worshipful order of bootstrap levitators.

Haha, but who knows how to write the compilers?

Who said there would be compilers in the future? That will become an OS thing, and “normal” programming will be performed using an abstract high level language using predefined libraries, and that the machine tweaks for better efficiency when the software is first installed.

This isn’t a dystopian future, this is what Android does today.

Essentially, we are doomed to creating fart apps, and huge (>20MB!) apps that reproduce the functionality of an existing website.

This isn’t a dystopian future, this is what Android does today.

Last time I checked, and it has been a while because I hate Android, it uses a Dalvik, a java-like virtual machine and then later “ART”. You need a compiler to compile Dalvik and ART so even if compilers become tools for OS and framework developers, someone will still need to know assembly, memory management and endianness.

Essentially, we are doomed to creating fart apps, and huge (>20MB!) apps that reproduce the functionality of an existing website.

Unfortunately that’s true and it’s hard to find a decent mentor to learn sensible programming as they’re outnumbered by bad programmers all publishing their own books/blogs. T.T

Unfortunately that’s true and it’s hard to find a decent mentor to learn sensible programming as they’re outnumbered by bad programmers all publishing their own books/blogs. T.T

If you expand the front of that to be “it’s hard to find a decent mentor that has the time to spare”

Like many people who really know what they are doing they are busy doing, meanwhile the one’s who don’t know and just talk rubbish are busy doing what they do best – talking rubbish.
Which makes it difficult for those of us wishing to learn to find the right person.

Like many people who really know what they are doing they are busy doing, meanwhile the one’s who don’t know and just talk rubbish are busy doing what they do best – talking rubbish.

So true!

Collected “wisdom”¹ from three generations (grandfather born 1890 with me as the sole name carrier now) and almost 6 decades of cynicism all my own. :)

Of course it may well go back further, but I never met my great grand parents.

¹ And zest for extracting the proverbial urine.

Essentially, we are doomed to creating fart apps, and huge (>20MB!) apps that reproduce the functionality of an existing website.

Alas, there are plenty of >20 MB websites out there too. It boggles my mind that a single page can need that much CSS/JS. Of course, my definition of “need” is significantly different from a modern “designer’s”…

We brought up a quick and dirty placeholder page at work (one of these “needs to be done yesterday” things) containing 20 words of text… and it uses 197 KB and doesn’t render properly in NetSurf. Sigh.

Alas, there are plenty of >20 MB websites out there too.

That’s why I left the web design. It was great when we had xhtml 1, but Google and Apple got in W3C and messed around developing a html 5 draft (which was unnecessary as XHTML 2 draft was great). I got annoyed when live websites were using HTML 5 before it was a standard and HTML 5 spec broke a lot of W3C standards reversing a lot of good XHTML 1 did. Suffice to say, I was sick of the web dev industry becoming so unprofessional by breaking accessibility of the web and decided to leave it.

We brought up a quick and dirty placeholder page at work (one of these “needs to be done yesterday” things) containing 20 words of text… and it uses 197 KB and doesn’t render properly in NetSurf. Sigh.

I remember aiming for 60KB max. That’s including images. It makes sense the number has risen because of higher resolution displays, but 20Meg is a joke. People don’t think about bandwidth much, anymore.

I remember aiming for 60KB max

A man after my own heart! I cut my teeth writing VIEW printer drivers. I was very pleased with what I managed to squeeze into 128 bytes of 6502 machine code.

I was very pleased with what I managed to squeeze into 128 bytes of 6502 machine code.

That’s very cool. 6502 computers are just before my time. I cut my teeth on an old IBM compatible Amstrad my Grandad gave me but didn’t really learn assembly until 12 years later!

Unfortunately that’s true and it’s hard to find a decent mentor to learn sensible programming as they’re outnumbered by bad programmers all publishing their own books/blogs.

Maybe contribute your own guidance?

Certainly as far as the RISC OS scene goes, there’s precious little in the way of developer documentation which isn’t rooted in the 1980s. What there is seems to be getting done largely by otherwise busy developers who recognise that there needs to be something out there to help newcomers.

Perhaps instead of being critical, you could write your own material or – more usefully, given the limited number of active developers available to do what needs to be done for the platform – offer constructive criticism to those who are currently working on stuff?