IT BS thread

I don't know if we're at the 'end' of Moore's Law per se, but I DO think we're at a plateau of sorts.

We've hit limits with CPU architecture. RAM can't speed up much more with current motherboards, and motherboards are limited by power, heat, resistance, speed, etc.

We'll need to get over the 'hump' with a breakthrough sometime fairly soon. Right now hardware is a bit stagnant and hasn't really changed much lately.

Think 8 bit to 16. 16 to 32. 32 to 64.

The next 'bump' will be to at least the equivalent of making 128 with exponential speed and performance increases the norm. And we'll need to start replacing a metric asston of hardware across the board to go with it.

Along with this will be infrastructure improvements / rollouts of ...well, something other than traditional copper to get the vast, VAST majority of the population on at least a rudimentary form of connectivity or broadband, bu that's a separate issue.

 

Right now, I really don't think we need more power. As a former programmer, I think we have gotten lazy and hardware capability has outpaced our ability to effectively utilize it. But that is just, like, an opinion, man.

 

It's kind of funny how faster CPUs and more memory makes a programmer lazy. Compiler technology has gotten better and programs aren't nearly as bloated as they once were but there's room for improvement in memory usage (in standard cell or memory development, it's not unusual to see the various tools used in the development consume huge amounts of memory for a very small result-to the point you see a 64G machine thrash.).

We use other tricks to do full chip simulations, like multiple processors with very large shared memories. I think you'll see more clever implementations like these. Anyway, lots of stuff to do to get past the need for bigger/faster CPUs.

 

some rambling thoughts:

I would agree with you. At one point I was tracking a think tank website (of sorts), where their sole purpose was to take old games and optimize past the point where most people would stop. Their one semi-famous achievement was taking Quake (or was it Doom?), adding an updated 3D engine (comparable to OpenGL, but strictly made for this program), fitting it on floppy with room to spare, and making it so that it would run without installation.

There's always more to be done in terms of code optimization. It used to be strictly a manual thing, and now it's more taken care of by the compiler.

My boss did some experiments with the SQL 2014 customer release and was seeing some pretty impressive gains.

 

Look at the first '64b' gaming systems. Nintendo 64 was 2x 32b R4300i's. Interesting way to solve the 64b requirement. At the time, 32b processor was easier and cheaper to develop.

Back to game systems. I think designers have to work pretty hard to make something that's efficient and cheap (there are Linux variants that will run on some game systems too). Pretty cool when you think about how they work-the same is true with modern graphics cards and their ability to function along side the system CPU.

 

I run into the lazy programmer all the time. Corporate world seems to take the cheap way out and throw a 20k piece of hardware at the problem to fix a poorly coded app. It would cost a lot more to make the app actually work right. They just keep promoting bad programming by buying a bigger better piece of hardware.

 

Hardware had indeed given shitty coders entirely too much to work with.

This isn't a 'get off my lawn' comment either. Back when you had limited memory to work with, and had to fiddle with the minutae of boot disks and config.sys/autoexec bat changes to free up an extra 5K, you had to have *tight* code.

Now? Meh, they throw CPU horsepower at it.

Wanna see some NICE code, even though I'm really sick of the guy that writes it? Steve Gibson.

On a related tangent: There is no program or app that should require a user to have god access to a workstation. Period. Bar none.

If you try to tell me that it does, it's poorly or lazily coded. I've been openly challenging people to prove me wrong on this for almost a decade and I have yet to find a single exception.

*rant off

 

You are correct. There are apps that require elevated privileges, but with the granularity that most OSs gives you for granting those privileges, There is not reason for "god" rights.

 

This is a bit long, but I gave a good belly-laugh to some guys here this morning and figured I'd share.

So, I'm filling in temporarily at a client site for a guy that's on paternity leave. High-buck law firm right on the water in West Palm Beach. I mean, DAMN. Here is the view from the IT desk:



Anyhow, I've only been here a couple of days, but...some times I can't explain why, but as my grandma used to say, "somethin' got my back hair up".

Just some random folks feel....off.

So it suddenly occurs to me as I'm walking out to the truck a little bit ago...a bunch of little details came together in my head. Lots of buzz cuts. Guys with teeny tiny waistlines and broad shoulders, really sharply pleated pants, NICE shoes. And the eyes!

It was explained to me a long time ago that, really, only two kinds of people will meet and hold your eyes directly in a random encounter: Cops and nutballs. Everyone else tends to look away quickly, politely, so as not to 'stare'. Not these folks.

So, I get back up on the 10th floor and I run into the office services guys in the lunch room, and I mention this to them.

One damn hear busted a gut laughing, the other was amused but at least would talk.

Turns out, the fourth floor of this building is DOJ leased and staffed entirely with ATF guys, and the sixth is an ICE/NSA office that doesn't exist.

I'm doing help desk in a spook central building!

(I look at it as a compliment that I noticed this only two days into my time here. They keep an ear out, and apparently there are folks that have been here over a year and still haven't noticed all the spooks with concealed weapons strolling around)

 

What is a spook?

 

Oh, sorry.

I picked the term up from my godfather as a yout. He would tell me of his time in the Air Force and local PD a lot.

Anyhow, spook was a general catchall term for non-uniformed law enforcement or federal agent of some type, to him. Essentially, armed and not armored as well.

 

Thanks

 

Entirely agreed, and the reason is that software is so intangible that you can hide the crap so long as you put a good chrome polish on it. It's part of the reason why I sometimes wish I had a different career: No one can tell if you really did a good job.
The top kid in my school could spit out garbage code at record speed and somehow make it do amazing things - no documentation, terrible structure, and about as clean with memory as a drunk with a full pint; yet it all would work in the end and would often do more than the assignment required. All that and typically a day early.
As for me, I didn't always finish. The program showed thought, but was basic to look at. However, under the hood I had it documented better than a textbook, code was concise, reusable, and structured like an engineer building a supercar.

Who gets the better grade? #1 is completion, #2 is end-user bling. No one can touch and see how well it's made since all of that is hidden or often too hard to understand by the layperson.

Fast forward to today, and it's the same thing in the corporate world. There's the classic example of the iron triangle, but I think it was put best by this commenter on this Coding Horror website post:

If you consider resources = better hardware, it still doesn't make the area (quality) bigger if budget and time are depleted.
I worked with an excellent programmer who left his job from burn-out. It stopped being about producing a product you could be proud of and knowing it would work for the customer, and instead became more about 'trash on time to maximize budget income'

 

I've designed systems that I am proud of conceptually (and the end result), but I had multiple programmers working on it. All of them had different styles. I had to get used to being less nit-picky, and as long as it stayed within the design and some level of standards, I had to accept it. As a system, it was great, but the under-the-hood details were not always great. There is time-to-market -- we wouldn't be doing this work unless there was a business reason for it.

 

If you want good code, give programers resources to quickly turn code (build machines) but make them run day-to-day and test on crap hardware-won't make them happy but they will become efficient

I am semi-serious

 

Anybody have need of a used but well-maintained Dell PowerEdge R210?



single quad xeon x3430 cpu
8gb 1066 ddr3 ram
250gb sata hdd
dual onboard gig nics
factory dvd

the usual

I've installed 2003, 2008sbs, 2012r2, and three different linux cores on it to test hardware and performance, everything works including all the fans.

Has a chassis-specific SBS2008 1-4cpu 5-cal license from MS included that you can use if you want.

Thinking...around 250, give or take shipping/delivery....I'm in zip code 33436 if you have a preferred way to go.

Figured I'd give first crack at it to y'all before sticking it up on Ebay

 

Morning guys,

Any of you every played with software called TeXShop/Latex? I have a user and he is getting an error in his code and I have chased it allover the web for the better part of the day. This same exact code works on his personal Mac OS MountainLion running TeXShop v2.47. He uses Dropbox to transfer the zipfile to the company Mac and it doesn’t work. The issue seems to be with line450 in the “PPT_october_13.tex” file. And getting error:

LaTeX Warning: File `PPT_october13-pics.pdf' not found on inputline 450.





./PPT_october13.tex:450: Package pdftex.def Error: File`PPT_october13-pics.pdf

' not found.
Any help would be great!

Thanks!

 

Totally off the top o me head:

First, confirm Mac version info between his machine and the company one, and not just OS. What version of PDF viewing software is he using on both ends? What file zipper/unzipper?

What application is zipping the app up versus unzipping? Is the dropbox transfer the issue or a non-issue?

Then, remove variables like dropbox. Does he get the same error if he doesn't zip the file, but instead brings it in via thumb drive or portable HDD, i.e. never haviong been compressed/uncompressed?

Etc.

Thinking out loud

 

Can you see the file on the destination machine? Is the path correct and accessible from the process running the script?

 

How is he compiling the file to begin with? Can you post line 450 of the source file? Also, did your user convert the original images to pdf?

The obvious part is to make sure that file (pdf) exists in the current directory. It's likely the graphics for his document.