Current version

v1.10.4 (stable)

Navigation

Main page
Archived news
Downloads
Documentation
   Capture
   Compiling
   Processing
   Crashes
Features
Filters
Plugin SDK
Knowledge base
Donate
Contact info
Forum
 
Other projects
   Altirra

Search

Archives

01 Dec - 31 Dec 2013
01 Oct - 31 Oct 2013
01 Aug - 31 Aug 2013
01 May - 31 May 2013
01 Mar - 31 Mar 2013
01 Feb - 29 Feb 2013
01 Dec - 31 Dec 2012
01 Nov - 30 Nov 2012
01 Oct - 31 Oct 2012
01 Sep - 30 Sep 2012
01 Aug - 31 Aug 2012
01 June - 30 June 2012
01 May - 31 May 2012
01 Apr - 30 Apr 2012
01 Dec - 31 Dec 2011
01 Nov - 30 Nov 2011
01 Oct - 31 Oct 2011
01 Sep - 30 Sep 2011
01 Aug - 31 Aug 2011
01 Jul - 31 Jul 2011
01 June - 30 June 2011
01 May - 31 May 2011
01 Apr - 30 Apr 2011
01 Mar - 31 Mar 2011
01 Feb - 29 Feb 2011
01 Jan - 31 Jan 2011
01 Dec - 31 Dec 2010
01 Nov - 30 Nov 2010
01 Oct - 31 Oct 2010
01 Sep - 30 Sep 2010
01 Aug - 31 Aug 2010
01 Jul - 31 Jul 2010
01 June - 30 June 2010
01 May - 31 May 2010
01 Apr - 30 Apr 2010
01 Mar - 31 Mar 2010
01 Feb - 29 Feb 2010
01 Jan - 31 Jan 2010
01 Dec - 31 Dec 2009
01 Nov - 30 Nov 2009
01 Oct - 31 Oct 2009
01 Sep - 30 Sep 2009
01 Aug - 31 Aug 2009
01 Jul - 31 Jul 2009
01 June - 30 June 2009
01 May - 31 May 2009
01 Apr - 30 Apr 2009
01 Mar - 31 Mar 2009
01 Feb - 29 Feb 2009
01 Jan - 31 Jan 2009
01 Dec - 31 Dec 2008
01 Nov - 30 Nov 2008
01 Oct - 31 Oct 2008
01 Sep - 30 Sep 2008
01 Aug - 31 Aug 2008
01 Jul - 31 Jul 2008
01 June - 30 June 2008
01 May - 31 May 2008
01 Apr - 30 Apr 2008
01 Mar - 31 Mar 2008
01 Feb - 29 Feb 2008
01 Jan - 31 Jan 2008
01 Dec - 31 Dec 2007
01 Nov - 30 Nov 2007
01 Oct - 31 Oct 2007
01 Sep - 30 Sep 2007
01 Aug - 31 Aug 2007
01 Jul - 31 Jul 2007
01 June - 30 June 2007
01 May - 31 May 2007
01 Apr - 30 Apr 2007
01 Mar - 31 Mar 2007
01 Feb - 29 Feb 2007
01 Jan - 31 Jan 2007
01 Dec - 31 Dec 2006
01 Nov - 30 Nov 2006
01 Oct - 31 Oct 2006
01 Sep - 30 Sep 2006
01 Aug - 31 Aug 2006
01 Jul - 31 Jul 2006
01 June - 30 June 2006
01 May - 31 May 2006
01 Apr - 30 Apr 2006
01 Mar - 31 Mar 2006
01 Feb - 29 Feb 2006
01 Jan - 31 Jan 2006
01 Dec - 31 Dec 2005
01 Nov - 30 Nov 2005
01 Oct - 31 Oct 2005
01 Sep - 30 Sep 2005
01 Aug - 31 Aug 2005
01 Jul - 31 Jul 2005
01 June - 30 June 2005
01 May - 31 May 2005
01 Apr - 30 Apr 2005
01 Mar - 31 Mar 2005
01 Feb - 29 Feb 2005
01 Jan - 31 Jan 2005
01 Dec - 31 Dec 2004
01 Nov - 30 Nov 2004
01 Oct - 31 Oct 2004
01 Sep - 30 Sep 2004
01 Aug - 31 Aug 2004

Stuff

Powered by Pivot  
XML: RSS feed 
XML: Atom feed 

§ Bloat

I just downloaded the latest trackpoint/trackpad driver for a laptop for a reinstall, and ended up with this:

[59MB mouse driver]

I don't mind a bit of expansion, but WTF... 59MB for a mouse driver??? That's almost as bloated as a printer driver.

I took a look at the expanded resources, and the big offender is that it includes the entire 43MB .NET Framework 4.0 Client Profile installer. Apparently it was worth dragging in the latest version of .NET in order to use WPF in the control panel, which, by the way, was already installed. On top of that, there are both 32-bit and 64-bit versions of executables that contain a ton of very large, uncompressed bitmap resources which presumably made up the bulk of the remaining 16MB of download. This took a while to download over a 1.5Mbps connection.

Is it too much to ask for just a little bit of attention to size?

Comments

Comments posted:


59MB? nVidia drivers are 147MB. Even with a lot of leeway--touchpad drivers don't have optimizing compilers embedded in them--the decimal point is still in the wrong place.

Glenn Maynard - 31 12 11 - 10:54


this is a sad trend from driver providers, i usually never download drivers from the laptop manufacturer themselves as they are always bloated, but rather locate the driver from the hardware manufacturer, in example; Intel Wireless driver from http://www.intel.com instead of http://www.hp.com, as intel usually provide 2 options, 1 bloated exe installer og 1 unbloated zip file with only .inf and .dll files.

And when i don't know what kind of hardware it is i copy the hardware id from device manager into pcidatabase.com

But it's not only drivers we are seeing this ugly trend, also in common applications like Office, Adobe Reader etc... Along the way when the harddrives increased their capacity the standards for coding also changed for compatibility and appearance over performance and stability :(

mehr - 31 12 11 - 11:20


The NVIDIA drivers were *200MB* in my case... but least I felt I was getting more out of that than from the mouse driver.

I remember once having to pull ATI's 26MB display driver over a 56K modem... that was really painful. I soon learned to grab the "core" driver instead.

Phaeron - 31 12 11 - 12:03


Check out the Synaptics trackpad driver - 55MB for 32-bit, 56MB for 64-bit one, and no .NET framework in sight (instead there's a bunch of .wmv files that demonstrate trackpad features in it's control panel).

And if we're speaking about huge driver packages, a recent ATI mobile driver is 318MB.

Luckilyy the connections here are fast, so the files download in minutes (well, except for stuff off Realtek's servers).

ender - 01 01 12 - 08:14


Well, that's quite natural, concluding the constantly reducing cost of storage/network bandwidth.... No one really considers that anymore, at least in the PC world. It's just the old-school-guys who gets annoyed by this nonsense....

bandana - 01 01 12 - 19:09


That's nice, except that you're ignoring that not all of those factors grow at the same rate in all technologies or even in different geographic locations, or that not everyone continuously buys the top of the line. Oh, and by the way, people with SSDs don't appreciate such waste given that they are currently about 7-8 years behind in $/MB compared to hard disks, nor do those virtualizing servers mind the extra delays in deploying VM images. But hey, feel free to pretend that the hardware guys can continue to fully compensate for inefficiencies introduced on the software side.

Phaeron - 01 01 12 - 19:27


My new laptop had a similar bloated touchpad driver (why can't they include the web-installer?), it even didn't work right with the default settings, I mean what's the point turning it off after some small inactivity time? Multitouch features are also a non-sense.

Gabest - 01 01 12 - 21:18


I agree with bandana, a 56 MB driver today is "cheaper" than a 50 kB driver was in times of internet over phone - both in terms of download time, HDD usage and memory consumption. Yes it sure is bloated but so is everything else including resources. And the thing is, not caring about size (and also not developing two driver versions, one for normal stupid users and one for old-fashioned-boys who know how to install a dll using inf file) makes the development cheaper.
And when it's about tight hardware restrictions, rest assured manufacturers make sure their drives fit. I don't really think anybody would need to install a touchpad driver on a restricted remote VM and even if yes, VMs can be deployed by tens to thousands from a single image and they don't care too much about dead weight the 95% of the installed system files is on the disk anyway.
For such kind of VM I'd not even consider Windows...

Kasuha - 01 01 12 - 22:42


And people, this is why interpreted/VM'd languages like .NET/Java/C# etc. are emphatically *bad*. This kind of bloat is simply unacceptable. No amount of "but if the client runtime is already installed, the actual code size isn't any different from native C/C++ code" hand waving will sway my opinion otherwise. Because you can't make that assumption. That's why they keep including the client runtime over and over again in download packages...

I do not buy the argument of managed code enabling faster development either. I just don't.

mpz - 02 01 12 - 01:21


56MB over broadband vs. 50KB over modem is off by an order of magnitude. A 28.8kbps modem transfers compressed data at 3KB/sec, so you'd need something like a 20Mbps+ broadband connection to match the 1000x size increase. Around here, that's a very fast residential connection.

By the way, guess what also hasn't scaled up as well: hard drive transfer rates. Here's a page with some nice graphs:
http://tylermuth.wordpress.com/2011/11/0..

Transfer rates relative to HDD size have also scaled by a lot... in the wrong direction. When I was on dial-up, I was getting sequential transfer rates in the 5MB/sec range. It's now around 40-80MB/sec. Not a huge improvement. It's even worse for non-sequential reads where seeking is a bottleneck and will continue to be so until SSDs are ubiquitous.

@mpz:
Interpreted/VM languages aren't the problem. The Lua runtime is lightweight (~100K) and Silverlight is relatively small given that it contains a miniaturized .NET CLR (CoreCLR) in 6MB along with everything else. The real problem is huge runtime libraries which can be a problem with native code as well: Qt's runtime is 10MB+. The argument that .NET was going to be cheap or free in size because it would be included in the OS was always B.S. simply because Windows can't ever ship fast enough to keep up with the newest versions of the .NET Framework, or just about anything for that matter.

Phaeron - 02 01 12 - 08:16


Ok, admittedly native code can suffer from that too, but at least MSVCRT is only on the order of a megabyte in size.

A trillion different versions of the d3dx libraries are a bit annoying too, but since the games they come with are often multiple gigabytes to begin with, it doesn't bother me too much.

mpz - 02 01 12 - 09:54


Internet speed is disputable - my local provider offers 10, 25, 50, and 100 Mbps programs so 20 can be considered 'near the low end' here. But that probably varies per area more than other things.

And you're right HDD speed did not scale up 1000 times since then.

The question is, are these two really worth complaining about? The delay on internet connection is in order of seconds, orders of magnitude less than time you need to locate the driver (which would take much longer over modem). And in case of HDD it's matter of approximately 1 second even if it was being loaded as a whole during each boot. Which I sincerely doubt as I believe most of those 50 MB will just sit on your disk without ever being used.

I used to complain about size of bloated programs too but I stopped noticing that as my internet speed and HDD size went up. I don't mind a few seconds wasted waiting for download or a few MB wasted on my disk. What I really dislike on these things are various side-effects they tend to have on my PC which are often hard to diagnose and fix (especially if they don't go away with uninstall).

Kasuha - 02 01 12 - 10:55


Ah, this brings back memories of when I installed vista and noticed that the install directory was over 20 gigs. I asked myself, how many lines of code could they possibly have for a 20 gig install?

Its not just windows either. Everyone is in this new bloatware routine. You download a program and they also want to install toolbars or other crap software with the package ( izarc, ccleaner to mention a few )

A relative program which everyone here uses such as xvid also started growing. Look at the size of this installer and ask yourself why?

Some other favorite things to notice are adobe installers for download. You have to download an installer, install it and then download the program you want when both files are the same size. WTF is up with that crap?

I now use places like freewarefiles to download programs such as google earch, adobe flash, java, etc to avoid the installers. I also look for portable versions to keep the bloatware down.

Bottom line, I can relate with Phaeron on so many levels on this issue. Internet speeds and HD sizes have nothing to do with the fact that people are letting programs grow without adding any value.

PS some other favorites are windows 7 installer directory, dot net install sizes and time for install and adobe products. My favorite example of them all is the hacked version of photoshop from 4 gigs to 4 megs.

evropej - 03 01 12 - 03:29


Kasuha - "Varies per area" indeed. PLEASE tell me where you live so that I can move there. Around here, the fastest connection available is 1.5 Mbps.

Tortoise - 03 01 12 - 10:51


@Phaeron:

I have ASUS P8Z68-V Pro mainboard which comes with Atheros Bluetooth adapter. Driver download for that piece of hardware is 322.28 MB in size!

If you unpack and analyze you will realize that ASUS has managed to bloat 50 MB manufacturer driver to 322.28 MB. True, they included both 32-bit and 64-bit install in one package, but they also put a lot of redundant stuff inside -- a whole other setup, duplicate files and all that so they can include two registry keys which they apply from small .reg file anyway.

It is all getting out of hand way too fast.

Igor Levicki (link) - 04 01 12 - 02:34


Bluetooth drivers have always been huge (and I still remember how a certain bluetooth driver searched for the My Computer icon, right-clicked on it, opened Properties and tabbed to the dialog where it could disable driver signature checking during install...). BTW, HP's Atheros bluetooth driver is 310MB, too (Intel's bluetooth driver is only 150MB).

And since we're talking about driver sizes, the (unpacked) driver for my RAID controller is 87068 bytes (and supports 17 different devices according to it's .inf file).

ender - 04 01 12 - 06:16


"And people, this is why interpreted/VM'd languages like .NET/Java/C# etc. are emphatically *bad*. This kind of bloat is simply unacceptable. No amount of "but if the client runtime is already installed, the actual code size isn't any different from native C/C++ code" hand waving will sway my opinion otherwise. Because you can't make that assumption. That's why they keep including the client runtime over and over again in download packages..."

That's how I feel about operating systems, just bloated libraries. I only accept code that executes in real mode.

fizzbin - 04 01 12 - 07:38


I suppose it was too much to ask for people to actually accept that others have differing requirements. I love how anytime there is an issue with performance, there is always an army of people available to insult. I'd love to see some of these people do technical support.

Phaeron - 04 01 12 - 15:16


Another point is that some people still pay per MB for their connections, and more and more people are moving to mobile connections too. This sort of thing adds up.

John - 04 01 12 - 17:08


Qt source tarball last I downloaded it was around 210MB. I can relate to 10+MB compiled runtime on that one beast. Then again I can stand behind py2exe packaging python interpreter into archive. Finding any version of python on windows machine is challange. On linux it's a little different story.

Iva9x - 05 01 12 - 06:31


HP must have included duplicate setup too.

Igor Levicki (link) - 06 01 12 - 11:45


Inside HP's atheros package are 3 separate installers for XP, Vista and Windows 7 (each is another self-extractor, which includes 32 and 64-bit self-extractors; this still leaves you with 50-60MB compressed installer for each platform).

ender - 07 01 12 - 00:10


If I am not mistaken, installers are all the same (universal) except for 32-bit .vs. 64-bit.

Igor Levicki (link) - 07 01 12 - 13:52


They aren't.

ender - 07 01 12 - 21:48


well, there is probably a bug per byte ratio too consider too.
a 130Mb pack of bit has probably a zillion more chance to incorporate a bug (either from the programmer or from an obsolete imported DLL or library) than a 50K file.

nosys70 - 08 01 12 - 08:54


This is a good laugh after a long day... My own tool has most the features you will find in modern GUI-based software, but is still only 600KB. I used to put user's manual in the build path to bloat it (to 3MB) so others will take it seriously. It is written in Java; so of course all libraries are pre-installed. But that's the point isn't it! Software getting bigger and bigger is simply saying people are re-inventing wheels all the time. I think VirtualDub deserves a hurray!

Yi - 11 01 12 - 04:53


I can't understand what's the problem with the bloated drivers. I do about 12TB internet traffic per year, so who cares about 10/20/50mb more?! There's no point in the whole topic, just go pay your internet and download as much as you want :).

WYSIWYG - 16 01 12 - 01:44


WYSIWYG
Even if you have a 20MB connection, the ISP limits maximum 250GB per month. No ISP that I know of has uncapped downloads. But that aside, you are missing the point. Bloating comes in many forms.

In five to ten years, you will need 0.5TB drive just to install OS and apps. Things are just growing from every angle and there is absolutely no reason why they should be other than lazy, ignorant, incompetent program software engineers from all walks of life.

I guess the bottom line could be that this is an indication of the quality of programmers and the standards of society. If people were getting smarter, things would shrink while adding features and speed. But who am I kidding, Intel and AMD cannot keep up with the bloatware exponential growing from all over.

This is exactly why I have some appreciation for the people who write a virus. Its tiny, it compromises security, its sets a server, its intelligent and its self sustaining. Give a Microsoft engineer the same task, and you will end up with a 500MB program.

evropej - 16 01 12 - 05:42


I have immense appreciation for the developers of uTorrent. It is a shining beacon in the sea of bloatware...

Too bad we seem to be moving further and further away from development practices that allow for compact and efficient code.

mpz - 16 01 12 - 15:21


"I suppose it was too much to ask for people to actually accept that others have differing requirements. I love how anytime there is an issue with performance, there is always an army of people available to insult. I'd love to see some of these people do technical support."

I'm sorry if you understand my comments as insults, they're not meant that way. I understand your feelings (actually I was feeling the same about bloatware for a long time) and actually I am also doing technical support too, although not kind you probably had on your mind.

What I'd like to say is, sometimes it's good to try to understand why is it so. Many people can get to a quite straightforward conclusion pretty fast - somebody is lazy or somebody is greedy. I believe the third common conclusion on such kind of matter - that it's a great conspiracy - does not come up in this particular case, but at least in my opinion they're all wrong.

What we see here in action is just normal evolution. Programs (and their programmers) must match selection criteria to survive and to have potential offsprings. If they don't match them, they die and are replaced by other, more successful species. And the selection criteria change over time as the environment where they "live" changes too.
Of course there are still programmers proud about creating fast and short optimized code. And there are still users which appreciate it. The problem is, writing such code takes way more time than writing something that barely works and is big. And people who do exactly that are in fact more appreciated because a program being bloated is not such a strong negative quality anymore and these people are appreciated for being more productive.

There's a great parallel in real life, too. Some people are amazed by the efficiency and geniality of how life works, storing information in DNA, copying, creating proteins which in turn support all that back. But in fact, life is the very first bloatware in the world we know. Taking just the storage aspect on mind, there are organisms with 1000 times more DNA than we have and they keep living even though just replicating all this DNA they never ever use takes a lot of energy. The evolution itself does not care about life being bloatware as long as it is competitive enough to survive in its environment.

Kasuha - 18 01 12 - 03:16


evropej I use AT&T, my Utorrent on my laptop (which is one of 3 computers that download files from the same connection) currently shows 450 gigs of files this month, yes since the start of the year. I haven't received any notice about a 250 gig cap. Admittedly I have downloaded alot more this month than most others, I know I have had other months where its approached similar amounts.

Flow - 18 01 12 - 13:31


@WYSIWYG:
Yes, we all should:

1. Pay for an unlimited Internet (even if we are on the road or out of the country)
2. Pay for the latest and greatest hardware to be able to store and unpack the thing
3. Pay for the electricity that keeps it running
4. Pay for the air conditioning to keep the damn thing cool
5. Pay for the coffee while we wait for it to unpack and install because no hardware will ever be fast enough to finish all the unnecessary and redundant crap someone thought it might be smart or cool to put in

Just because someone, somewhere cannot bother to either:

a) Write a touchpad configuration application and driver without using .Net framework

or at least:

b) Offer driver download package without .Net Framework just in case you already have it preinstalled so you don't have to re-download those 43MB every time you update your driver

When we are at it, even the 16 MB for a freaking MOUSE (because that is what touchpad is in its essence) is two orders of magnitude too much. That is coming way too close to the size of an OpenGL ICD which implements full OpenGL specification while MOUSE driver just reports X/Y coordinates and button clicks.

@Kasuha:

By your logic all those people getting fat and lazy are actually "evolving" into superior beings. I am glad that I am not the only one who disagree with you.

Igor Levicki (link) - 18 01 12 - 21:59


@Igor Levicki

In evolution, the most successful species are not those which you are rooting for, rather those which are best suited for their environment.

And notice that it's not my fault, it's a general principle. You may of course disagree with me but that does not change anything on it.

Kasuha - 20 01 12 - 18:38


This make me I recall an article name Small is better of Steve Gibson.
And all those stuff now become huge.

Viet Le - 27 01 12 - 19:06


@Kasuha:

If some organism has extraneous DNA it doesn't mean that DNA is useless -- it is more likely that we didn't figure out it's purpose yet.

The same cannot be said for software -- bloated code is a byproduct of inferior code production process, and it cannot serve any other purpose.

On the other hand, DNA even if it seems extraneous at some point, can probably be repurposed in future.

Therefore, your analogy is completely flawed.

Igor Levicki (link) - 28 01 12 - 03:50


I recently came across something similar: I work with z/OS XMIT files and needed something to extract them on the PC. Found a Java based tool on SourceForge, a snip at just 56Mb. Then someone pointed me to a simple exe, which comes in at a hefty 39k (yes, THIRTY-NINE *kilo*byte) and does the same thing, be it from the commandline...

Of course there is also the seminal page @ http://www.ncdm.com/bloat/bloat.htm

Robert (link) - 29 01 12 - 04:04


All you guys complaining about downloading these bloated pieces of $#!^ over your measly 10/20/50/100 Mbps connections, think what we have to deal with, given that we pay through our noses for a 'high-end' 1 Mbps connection (that actually gives us 30-40% of rated speeds at most), with 'fair usage' policy rubbish, bandwidth caps and what not. We have the same hardware (for which we pay a premium too of course), so downloading those same drivers reminds me of downloading 1 MB+ files over dial-up. At least these damn connections don't disconnect. I can remember wearing a hole in the floor pacing up and down worse than a father-to-be, waiting for a file to download before the connection snapped. Ah, what a blessing GetRight was when it first appeared on the scene! :)

Karan - 29 01 12 - 22:35


I think, as hardware improves, it becomes natural for programmers to get a bit lazy, and to a degree it's acceptable.

For instance, in the old days, to get the most out of hardware, some important loops of software, or important functions such as calculations, had to be written in assembly code, writing them in a high level language like C and compiling them, wasn't fast enough, and made the code too bloated. Now though, we're at the point where hardware is fast enough to run most software in a VM with a language like Java. Which I guess one could say is 'lazy', but it's still quite acceptable really.

.. (wait for it).... but!! BUT!!!

That is NO EXCUSE, for adding such insane bloat to software, as we see today happening all over the place!

I remember back in the days of Windows XP, it was reasonable for the OS to boot up and load, showing a desktop, with a start menu, with windows, and networking, and all that good stuff, all within a few 100mb's. Now, my computer eats up 2GB's of RAM just BOOT. For what? What's so amazingly different about Windows 7 that it has to eat up 20x more RAM than Windows XP? The user interface is a little prettier, but it's just a freaken shell, that can't be what's taking up the extra space!

Firefox right now, I just checked, is using 2GB's of RAM on my computer! Thank god I have 16GB's of RAM :L

Computer games?... It's insane, some games take up 15-20GB's of space, and if you install their expansion packs, some of the expansion packs don't even use the currently installed version of the game's files, and just take up the same space again! GTA:4 + GTA: Episodes from Liberty City = 30GB+!

A tonne of programs I use now include things like AOL Toolbars, Ask Toolbars, and other junky addon software, which is all just adware and junk, but for some reason included in the installer. Most of it would be bloating the installer out to a size at least 10x what it needs to be. Not to mention all the software which installs it's own version of the .Net Framework, to add to the 20 different versions I already have installed..

I see a bunch of programs now using uncompressed high resolution animated resources, blowing out the size of the application with little fancy effects that add no value to the application at all.

Driver installers, which have many great examples in this post, are another good example.

OS installations, like Windows 7, can take up 10-20GB's space, when they should only need 3-4GB's at most!

There's being lazy because computer hardware has advanced enough that you don't have to be anal about every detail.. and then there is being lazy for no good reason at all, in ways which don't even make sense. As applications like uTorrent demonstrate, the massive increase in file size for applications isn't necessary, and application developers just need to lift their game. Another good example is Ninite.com.

Losing 5% performance in an application, by taking a shortcut that saves you 95% of your development time, is perfectly reasonable. But losing 90% performance by taking a shortcut that only saves you 5% your development time, IS DEFINITELY NOT.

Using 5% more hard drive space, by taking a shortcut that saves you 80% of your development time, is perfectly ok. But using 500% more disk space by taking a shortcut that saves you only 20% of your development time, IS UNACCEPTABLE AND A FAILURE.

BTW, file sizes for things do matter to me!

In my country (Australia), the most common connection speed is around 1mbps, and when designing websites, for users I have to allow for as low as 256kbps, because those speeds, believe it or not, are actually still very common in this country (50GB download+upload limits are common too). So it's not a difference of seconds to us, it's a difference of minutes and sometimes even hours. Personally, I'm a bit ahead of the curve with a 14mbps connection, with a 200GB download+upload limit. But even to me, downloading a 30mb driver installer instead of a 300mb driver installer, is HUGE difference to me (17 seconds vs 3 minutes! ) and even bigger difference to other folks in this country! (A 300MB file with a 20mbps download speed, is a comfortable 2 minutes. But for someone with only 256kbps speed, that's over 2 HOURS!)

Not only that, but my computer has a SSD disk, it's only a 240GB SSD. That works out to 222GB's as Window's reads it, 15GB's of that is just Windows, and the software/drivers/AppData comes to 96GB's. Add another 100GB's to that, which is just my personal files (videos and stuff are stored on an external drive, they just won't fit on this machine). Right now I got about 5GB's of free space on my hard drive. So bloat is important to me, I recently changed what office suite I use because the one I had was 1GB in size, and the new one was only 300MB's. So bloat isn't just a thing I whine about, and go 'Oh how unprofessional!', then keep using the software.. Bloat for me is something that's very serious. If an application is too bloated, I try my best to find an alternative that will do a reasonable job of replacing it, so that I can cut down on the hardware usage. I switched from Vuze to uTorrent mainly for that reason.

So I guess what I'm saying is.. "Is it too much to ask for just a little bit of attention to size?". Nope, not at all, more software developers need to pay attention to the size of their software, and just how many corners they are cutting and what it's costing them. Not everyone in the world has top of the line hardware, or high speed/limitless/low ping internet connections. Besides there are benefits for the developers too. Surely it would be less of a load on the Dell/ASUS/AMD/Intel/Etc servers, is the drivers were 1/10th of the size they are now when being downloaded.

Grady - 31 01 12 - 00:53


Well, just because it *appears* that Windows 7 uses 2 GB at boot, doesn't mean that it actually requires as much. Unused RAM is wasted RAM, so even at boot the OS will use memory for disk caches and preloading software that you're probably going to launch soon.

http://blogs.msdn.com/b/b8/archive/2011/..

If you only have 1 GB of RAM, Windows 7 SP1 will only use 400 MB at boot. Windows 8 will further reduce it to 280 MB. They also point out that the graphics driver is a relatively big memory hog these days (it didn't use to be), so by unloading that you will get the memory usage below 200 MB in Windows 8. That means the core OS requires less than 200 MB, and I think that's a pretty damn low number for an OS from the year 2012.

But yes, HDD space tends to be wasted for little to no reason these days.

mpz - 31 01 12 - 19:07


Personally, I think that including disk cache in the memory total is misleading at best. I'm not sure why that is useful since it's memory that (a) is very easily reclaimed and (b) is speculatively committed by the OS instead of by the applications themselves. It's certainly much less impact than swapping out code or even worse modified data.

My experience has been that Windows itself isn't too bad memory-wise, especially if you exclude the memory hungry DWM. Applications dwarf the OS in memory usage very quickly. Of course, the fact that Windows is much more memory efficient when you aren't doing anything is moot.

The disk cache and preloading, btw, is the main argument I use as for why applications shouldn't feel free to use otherwise free memory just because it's there. You're not the first one who had that idea, and the OS disk cache might use it more effectively than you.

Phaeron - 31 01 12 - 19:21


Agree with pretty much everything Grady said, except for SuperFetch in Vista+ leading to high RAM usage levels. Sadly, uTorrent is another casualty of bloat now. Recent versions (the 3.x line especially) have started including all sorts of crap that has absolutely nothing to do with the core functionality, which is to simply download torrents. I wish just like their uTorrent Plus paid version, they'd go ahead and introduce a uTorrent Lite. I'm sure *lots* of people would be interested in the same.

Karan - 01 02 12 - 22:23


Hereby I propose new term for bloat -- code cancer.

Igor Levicki (link) - 02 02 12 - 02:12


Slightly OT, but why does the frontpage still say that there's no comments? :)

ender - 04 02 12 - 04:52


I have to agree with this. It seems like nobody mind wasting the ressources of customers.
But if you think 59mb is in the range of printer drivers, you don't have seen HP drivers. It's pure madness to have several hundred mb for an automatical update manager.
If I remember correctly it's about 800 or 900mb. As big as the whole rest of the driver 'suite'.
Even without this my HP printer driver is the biggest non OS or Game software i have on my computer.

But it's really nice to see that some people also think thats kind of crazy, not saying 'what? just buy better hardware'

Bigeagle - 07 02 12 - 04:25


60MB for a simple little driver. wow

i was watching the old Demo/Intro "theprodukkt" http://www.theprodukkt.com/demoscene and am amazed how much that can fit in 64KB, and some people cant even make a decent trackpad driver.

Simon - 14 02 12 - 06:14


And of course there are the anti-virus software vendors.

Every other day another 15 to 20 megabytes of 99.9% the same as the previous update. And to make it even worse their poxy overloaded servers trickle the stuff out sometimes at bytes be second.

IanB - 14 02 12 - 07:52


Adobe, HP and others adds hundreds of different language files, when it should only copy the english and the local version instead. Adobe is the worst, I found 5gb in 45000 files of localization files using advance search (ie: ar_ae|ar_sa|az_az|...) in Total commander. Apple software takes more space of localized DLLs, helps, etc than the rest of the application. It's stupid. Besides Adobe software does not fit on a single layer DVD anymore.

isidro - 26 02 12 - 18:51


How the fark do you burn 5GB in localization resources? Rasterize every single dialog in every language in HDR at 4Kx4K and store it as PSD?

Phaeron - 27 02 12 - 16:49


Maybe they included text-to-speech as audio files for each language? :)

ender - 02 03 12 - 22:03


What bothers me the most is that I don't use the trackpoint. I prefer the mouse so I have to install this bloatware just to disable the trackpoint just because Dell can't figure out how to disable it in the BIOS! Stupdid!!!
BTW, I have a Dell XPS M1530. One of the best thing about this laptop is it's excellent sound (with earphones) and a good low noise line-in connector. Not just a microphone connector.

Snickie - 14 03 12 - 05:15


Time for a Windows package manager.

Balian - 16 03 12 - 19:45


Balian: CoApp?

ender - 17 03 12 - 02:56


No, they include folders with these names: ar_ae|ar_sa|az_az|ba_ba|be_by|bg_bg|ca_ca|ca_es|cs_cz|cz_cz|da_dk|de_ch|de_de|el_gr|en_au|en_ca|en_gb|en_xc|en_xm|es_as|et_ee|fa_ir|fi_fi|fr_be|fr_br|fr_ca|fr_fr|fr_xm|gl_es|he_il|hi_in|hr_hr|hu_hu|is_is|it_it|hy_am|id_id|in_in|is_is|it_it|ja_jp|jp_jp|kk_kz|ko_kr|lt_lt|lv_lv|mk_mk|mk_yu|mt_mt|nb_no|nl_nl|nn_no|no_no|pl_pl|pt_br|pt_pt|ro_ro|ru_ru|sh_yu|sk_sk|sl_si|slo_slo|sq_al|sr_cir|sr_Latn_CS|sr_yu|sv_se|sv_sv|th_th|tr_tr|ua_ua|ug_cn|uk_ua|va_es|vi_vn|zh_cn|zh_hk|zh_mo|zh_sg|zh_tw
And ITUNES/QUICKTIME (>100mb): French.lproj|German.lproj|Italian.lproj|Japanese.lproj|da.lproj|de.lproj|en_GB.lproj|fi.lproj|fr.lproj|it.lproj|ja.lproj|ko.lproj|nb.lproj|nl.lproj|pl.lproj|pt.lproj|ru.lproj|sv.lproj|zh_CN.lproj|zh_TW.lproj
French.lproj|German.lproj|Italian.lproj|Japanese.lproj|da.lproj|de.lproj|en_GB.lproj|fi.lproj|fr.lproj|it.lproj|ja.lproj|ko.lproj|nb.lproj|nl.lproj|pl.lproj|pt.lproj|ru.lproj|sv.lproj|zh_CN.lproj|zh_TW.lproj

They have each DLL and so on compiled for each language in multiple folders, I also saw dictionaries for all languages. What's worst now is that Adobe Acrobat detects missing files and restores all the multilanguage garbage again.

isidro - 20 03 12 - 05:27


HDD speeds sure haven't increased at the same rate their capacity does, but they are getting faster. Some of the new ones are inching towards 200MB/sec sequential. Even mine from 2 years ago peaks at 155MB/sec. (avg 120) Read caching helps to cut the I/O down, which is one of the reasons I have 16GB of RAM... but write caching helps far more. Software like FancyCache or SuperCache can cache writes to RAM - it's insane how much it improves performance. Tasks which took several minutes before drop down to dozens of seconds. It's as fast as having a high end SSD.

The only problem with software like FancyCache is, one BSOD and you have massive filesystem corruption. I don't like that.

But hey, if you're like me and think drives need more cache, vote for it:
http://community.wdc.com/t5/Internal-Dri..

On-drive cache isn't vulnerable to BSODs. The cache still drains to HDD platter after your system has locked up. More cache allows drives to be more efficient by combining reads/writes into bigger sequential ones. It definitely has a positive effect on performance, so I don't know why cache sizes are increasing so slowly.

Kramy - 25 03 12 - 11:46


I feel the pain of people stuck on slow connections. I'm on 6mbit right now - magnitudes better than 1.5mbit, but when I'm waiting for a Realtek audio driver to finish at 30KB/sec, max connection speed really doesn't matter. :P

I'm in western Canada. I guess the one positive that we have out here is none of the western ISPs have enforced caps. Our cable provider offers 250mbit now. A few people were posting on forums that they downloaded 20TB/mo for 4 months straight without any peep from their ISP. (One guy had a 48x3TB RAID array.)

Kramy - 25 03 12 - 11:53


My 4 year old acer laptop bluetooth driver was 71mb but windows automatic search was able to find a driver without any bloated utilities etc for less than 5mb.

Lakshmi Rai (link) - 02 04 12 - 21:19


I've downloaded network drivers from Belkin and Linksys that included Acrobat Reader. At least one driver download included two different versions of Acrobat Reader. A driver download should include the *driver only*.

Galane - 08 04 12 - 21:43


I agree.
But personally I reject anything that requires .NET.
It's a horrible unit that never completely uninstalls, and does weird slowdown stuff on my machine.
BTW your front page sccript appears to be broken. The post showed 'No Comments'.

Matt (link) - 15 04 12 - 19:38


WiFi driver update for my Lenovo E320 is 257 MB. I have no idea why.

Artifex Maximus - 26 04 12 - 20:36


I wrote a typewriter AND a concurrent alarm clock in machine code using just 2k bytes. That is 1024 bytes in case you guys have forgotten those little things! These programmers should be sent back to the early days of computing. 32k and you could run a C compiler and editor and linker. Indeed C was written on a machine with less than that. You couldn't even get ONE of these bloatware bitm.aps into that!
I despair

Jon Heather - 03 05 12 - 01:59


not a big deal. mostly you can browse the .msi or .exe with 7zip, extract what you like(the driver folder) and there you go. mostly these files include a common installer(similar in behavior to batch files) and standalone setup.exe-s(or whatever the name might be)in their respective folders. You might just extract the touchpad driver folder and run the standalone installer inside of it.
Aftet the install you might just resize the .bmp images or even delete them if they aren't any important icons-programs load em but hardly ever check or report errors if they're missing.
the same way you can get .inf files for nlite vlite 7customizer and sort.
as for nVidia-it extracts to a temporary location (defaults to c:\nvidia -set it elsewhere to retain system partition speed-the first option-it's not the install loc. itself) then save what you need elsewhere for future reinstalls or use the infs for your custom install disks.

dushko - 13 05 12 - 20:37


One time I grabbed a printer driver for a Kodak printer, fecking thing was half a gig.

Burito - 19 05 12 - 05:53


"Is it too much to ask for just a little bit of attention to size?" Yes, I think so. It's just beyond many people. Few seem to have any concept of the size of things "you know, digitally". I routinely get Word documents from companies and government departments that are many megs in size, yet half a page. Turns out their logo is some ultra high res prepress thing, inserted completely uncompressed somehow (Word usually compresses pasted images as PNG at least), and these places have IT experts. You'd think a manufacturer or SW company would know better, but when you think about it they're no different.

Antony - 19 05 12 - 20:36


It's refreshing to see somebody relatively important speaking out against this degenerate trend in software. It's appalling how many fucking morons are minimizing or denying this problem in spite of the overwhelming evidence. I attempted to bring this up with the Firefox devs who had their acolytes assure me the reason Firefox drags on my modern i7 computer is because it's that fucking awesome that it requires 20x more of my resources.

I must be tripping balls, because I've struggled to notice what's so awesome about the new Firefox and still can't. Apparently, their huge loss of market share to Google Chrome can't possibly having anything to do with their lack of competency. Hahahaha, what retards.

I remember the days before any trailer-dwelling retard could figure out how to log on the net. The resources were very limited but if you knew your shit and put some effort into efficiency and innovation, the experience could be very rewarding. It taught you modesty and respect. And when that new 128MB RAM came out, it was a major relief as you didn't have to be so conservative about resources and could open your mind to further innovation.

4GB would've been a fucking dream come true. To have active memory the size of your hard disk... damn. Everyone thought we'd have Johnny Mnemonic OS's and interfaces with that much abundant memory. Little did I or everyone else think it would somehow be wasted by the same interface with the same features with the same if not LOWER level of convenience.

I am truly at a loss of words. Since when did "barely functional" become the minimum standard for everything?

Xpenguin17 - 10 07 12 - 08:33


With very few exceptions, all drivers now require extraction and sorting out the driver from applications. It's not as easy as to unpack a zip file. Many drivers use self-extracting installers from InstallShield or custom ones, which require advanced tools to unpack or another machine where the installer can be run safely, without contaminating the system with apps.

This is a lot of wasted time. I've recently finished finding drivers for a laptop that was brought in for repairs. I didn't find a manufacturer's site listing all the drivers for this particular model, and had to download quite a few large exe installers and unpack them to find matching drivers.

I am keeping an archive of all extracted drivers, with packages renamed to an uniform scheme, which includes the manufacturer, model series, version number and operating system(s). This saves time in the long run especially for common devices such as network adapters. Extracted drivers are usually quite small, and sometimes multiple OS versions can be compressed in a solid archive well.

Drivers definitely are not the place to include localized strings, because the parameters translated are very technical, and only understood by systems administrators. Yet, translations occur in INFs, which are then usually also converted to Unicode doubling their size.

Driver availability for old systems, since everyone is pushed to perpetually upgrade and update, is poor. No matter what my or yours network connection is, it obviously is expensive for groups to keep old drivers online. That could have been remedied if drivers were smaller.

I was on the market for a cheap printer recently and picked a Canon. I must have made the right choice, because, unlike HP, their drivers are small.

Extracting drivers as I descrived worked well for a decade. But what does nVidia do now? Their drivers won't install anymore from INF files, because the process is blocked by a failing custom Co-Installer. If I take out the coinstaller from the INF, the driver doesn't include any control panel, and becomes unsigned. To get the (essential) control panel, or parts of it, one has to figure out how it is registered on the system.

I can only speculate as to why they would so aggressively push that the whole package be installed. But it is obvious to me that that is happening. It defies logic that an office machine, or a media center in the living room, would need the physics drivers or the control application *after* the right options have been set on it (such as video color settings, because it the driver defaults to Full Range input instead of Studio for YCbCr). The preview graphics included in the bloated control panel are not even remotely accurate, and quite misleading. For example, the Anti-Aliasing page includes a heavily sharpened picture for anti-alising Off.

The stability and features of nVidia Fermi drivers are alright. But the driver package has highly disappointed me.

j7n - 03 03 14 - 06:44

Comment form


Please keep comments on-topic for this entry. If you have unrelated comments about VirtualDub, the forum is a better place to post them.
Name:  
Remember personal info?

Email (Optional):
Your email address is only revealed to the blog owner and is not shown to the public.
URL (Optional):
Comment: /

An authentication dialog may appear when you click Post Comment. Simply type in "post" as the user and "now" as the password. I have had to do this to stop automated comment spam.



Small print: All html tags except <b> and <i> will be removed from your comment. You can make links by just typing the url or mail-address.