Just a question which you might have already considered; are those companies willing to support this product for 10 years? Bear in mind that any enforced upgrades (i.e. we won't support version 2, you have to upgrade to version 4 or we withdraw support) will cost time in terms of manpower required to ensure a smooth transition and possibly higher hardware requirements. Using an OSS solution means you can stick with whichever version works for you and patch as required.
Added to this, once you've bedded the product in, probably over a course of 2-4 years, it will just sit in the background and work as-is without any intervention.
Hrm; when I was at uni, I was able to walk into a bookstore and buy MSVC 5 + VB 5 for £100, far less than RRP. All I had to do was have my lecturer sign a form saying that yes, I was a student. Sent that off to MS and got my software back.
This was, obviously, before I became so cynical (some might say englightened) of MS.
At the time, I don't think my uni had an agreement with MS, that only came in about 3-4 years after I finished (I started working there, so I know when it happened).
Re:On the matter of vendors
on
Nosy Vendors?
·
· Score: 2
Very good point on focussing on your needs. When you go to a vendor, decide the following:
What you're trying to achieve
How you're going to support the solution
Your budget
There's no point speccing up a gee-whiz bang solution with all the industry buzzwords and acronyms if that's not what you need.
Start with the problem you're trying to solve, and work towards the solution. Don't let salesmen railroad you to their preferred sale.
Finally, salesmen have one goal; to sell you stuff. Preferably high value stuff which gives them bigger bonuses/commission; some will lie & cheat to a greater or lesser degree to achieve the sale.
The maintenance nightmare is the best point; why spend x man-hours writing a custom data store when there are thousands of man-hours of experience in Oracle/PostgreSQL/MySQL/MS SQL/whatever? You have access to tried & tested code under a huge variety of circumstance.
Take the time saved by using an out of the box solution (i.e. the DB) and buy a bigger box to handle any inefficiencies you get by using a DB.
That's a good point; much as I'd love to add stuff to my web pages, I don't want to block out some of the lower denominators such as lynx or, possibly more importantly, software such as readers for the blind. If DHTML screws up on those, you're losing a portion of your audience; not perhaps a large one, but it's still there.
The reason is simple; it means you can have the same OS on your cheap x86 clients as your high-end SPARC servers.
BTW, my understanding is that most of the complaints about speed are due to two factors;
Solaris isn't as friendly in low-memory machines as linux; it's optimised for n-way servers.
Solaris' IDE support stinks (or at least it used to; Solaris does now support DMA if you poke the right config files); from what I understand if you run it on decent SCSI, you'll do fine.
Depends where you shop; eBay usually has some cheap, and I've seen Sun 21" monitors for about £300 before. These are nice monitors, with dual inputs as well.
I've had a few problems with mine overheating & locking up (non-overclocked Athlon 1800+, 1.53GHz), usually while trying to rip an album to MP3 using grip/lame. I'm not sure if it's down to dust collecting in the fan/heatsink or poor thermal conductivity. However, removing it, applying another layer of thermal grease and re-attaching has fixed the problem twice for me, but you shouldn't have to do that. The next time it breaks on me, I'll have to get a new batch of grease, so I'll try a quality brand (I think it's some fairly cheap stuff supplied with mine).
I wish more systems had something to control the speed of the fans. As an example, some newer Sun kit (e.g. Sunblade 1000/2000) have a process which monitors internal temperature and adjusts the fan speed accordingly (NB: if the process dies for any reason, the fans start going at full speed). Ultra 30s/60s have an add-on kit to reduce fan speed at low temperature.
Having just fitted the noise reduction kit to my Ultra 30, I can confirm it makes a hell of a difference to noise levels. The total of the kit is:
A temperature sensor
A voltage regulator (based on the temp sensor, I assume)
two fan outputs
a temperature sensor
a "through" cable for the power
It didn't cost that much ($35), and I'm pretty sure it could be made cheaper if integrated into a motherboard. Normal users wouldn't even need to know it was there, as all the above parts would be built into the mainboard and the fan outputs would adjust voltage levels as required.
While they probably won't profit share, they'll likely share in other ways, by code improvements etc. IBM is investing heavily in linux and I'd assume they're looking at ways to improve linux to make it as stable as AIX is. They've already done work on integrating JFS into the kernel, for instance.
Well, let's compare this to other desktops; XP allows me to do all sorts of stuff with the user I created at initialisation, including create other user accounts. What this is probably aiming for is the same sort of user as XP, i.e. one that doesn't want to know about 'root' he wants things to 'just work'. In any event, it's probably more secure than having the user log in as root with a password of 'password'; how long would that take to crack?
If you're going to deploy this in any kind of serious setting, you'll have admins to set up scripts to remove the glaring security holes in any case.
I think most of them run AIX on IBM hardware; not sure what proportion.
This raises another question; of the ones which survived, which OS's were they running? It would be interesting indeed if the only surviving name servers were the ones running a specific OS...
My guess is that it's probably technically possible, but will require that KDE/Gnome/whatever supports the modification of the X-server when you switch desktops.
It does sound like a very good idea and I hope it gets implemented.
Which is fine, if you can keep enough airflow in the closet. If you can't, you'll just keep generating heat in that closet and it'll start to overheat.
The test pilots will be highly skilled (you want the best on this kind of job) and probably don't need flight computers for anything other than complex navigation or flying in thick fog/cloud.
Boeing wanted a rapid and fairly inexpensive development cycle; this is, after all a prototype.
Computers need to be tested when you put them in aircraft, especially experimental ones where you have to throw out much of your previous learnings.
In an experimental aircraft, the computer programming would have to change with each iteration of the aircraft; this 'tweaking' could cause bugs to creep in and would certainly add to the time required to create a new version of the craft.
Based on that, adding flight computers would have been expensive, time consuming and wouldn't add to the value of the experiments. It would also have added weight & power consumption to the craft, neither of which is desirable. If it were to go anywhere near production and real use, that is when you start looking at the computers.
With the obvious problem that you're spending $4000 to back up a system which probably cost
Given the cost of DVD burners + media, you're considerably cheaper looking at DVD for backup than any tape technology.
I was thinking about this earlier, and it's nothing I couldn't do with Samba (smbclient) and a short perl script (heck, even ksh could do it).
As for people saying "turn off the messenger service", there are actually valid uses for winpopups. At my last work, I set up a few perl scripts that would use smbclient to warn Samba users when they were over quota. Before that, users would go over quota and wouldn't know about it until things broke after the grace period.
Obviously, you should be filter Netbios ports at the firewall unless you have a damn good reason to have internet access to them. If someone in your network is using this program to spam, the LART them appropriately.
Slashdot Mirror
Added to this, once you've bedded the product in, probably over a course of 2-4 years, it will just sit in the background and work as-is without any intervention.
That's what Sun thought with Java...
For some (l)users, their background picture is very important...
This was, obviously, before I became so cynical (some might say englightened) of MS.
At the time, I don't think my uni had an agreement with MS, that only came in about 3-4 years after I finished (I started working there, so I know when it happened).
- What you're trying to achieve
- How you're going to support the solution
- Your budget
There's no point speccing up a gee-whiz bang solution with all the industry buzzwords and acronyms if that's not what you need.Start with the problem you're trying to solve, and work towards the solution. Don't let salesmen railroad you to their preferred sale.
Finally, salesmen have one goal; to sell you stuff. Preferably high value stuff which gives them bigger bonuses/commission; some will lie & cheat to a greater or lesser degree to achieve the sale.
Take the time saved by using an out of the box solution (i.e. the DB) and buy a bigger box to handle any inefficiencies you get by using a DB.
That's a good point; much as I'd love to add stuff to my web pages, I don't want to block out some of the lower denominators such as lynx or, possibly more importantly, software such as readers for the blind. If DHTML screws up on those, you're losing a portion of your audience; not perhaps a large one, but it's still there.
Kudos to the Ars team yet again for going deeper into CPU designs than 99% of the populace need to go :)
BTW, my understanding is that most of the complaints about speed are due to two factors;
Depends where you shop; eBay usually has some cheap, and I've seen Sun 21" monitors for about £300 before. These are nice monitors, with dual inputs as well.
I've no idea what happens with the fans if a processor dies.
I've had a few problems with mine overheating & locking up (non-overclocked Athlon 1800+, 1.53GHz), usually while trying to rip an album to MP3 using grip/lame. I'm not sure if it's down to dust collecting in the fan/heatsink or poor thermal conductivity. However, removing it, applying another layer of thermal grease and re-attaching has fixed the problem twice for me, but you shouldn't have to do that. The next time it breaks on me, I'll have to get a new batch of grease, so I'll try a quality brand (I think it's some fairly cheap stuff supplied with mine).
Having just fitted the noise reduction kit to my Ultra 30, I can confirm it makes a hell of a difference to noise levels. The total of the kit is:
- A temperature sensor
- A voltage regulator (based on the temp sensor, I assume)
- two fan outputs
- a temperature sensor
- a "through" cable for the power
It didn't cost that much ($35), and I'm pretty sure it could be made cheaper if integrated into a motherboard. Normal users wouldn't even need to know it was there, as all the above parts would be built into the mainboard and the fan outputs would adjust voltage levels as required.While they probably won't profit share, they'll likely share in other ways, by code improvements etc. IBM is investing heavily in linux and I'd assume they're looking at ways to improve linux to make it as stable as AIX is. They've already done work on integrating JFS into the kernel, for instance.
If you're going to deploy this in any kind of serious setting, you'll have admins to set up scripts to remove the glaring security holes in any case.
I'm sure people will, once the poor server has recovered from the stampeding hordes of /.-ers who have just pounded it into the ground.
This raises another question; of the ones which survived, which OS's were they running? It would be interesting indeed if the only surviving name servers were the ones running a specific OS...
SearchKing, Oklahoma's premiere parasitic link-farm
vs
SearchKing is one of the pioneers in developing portal and search engine software and services
Guess which report has which statement? :)
It does sound like a very good idea and I hope it gets implemented.
Which is fine, if you can keep enough airflow in the closet. If you can't, you'll just keep generating heat in that closet and it'll start to overheat.
$ make love
Not war?
- The test pilots will be highly skilled (you want the best on this kind of job) and probably don't need flight computers for anything other than complex navigation or flying in thick fog/cloud.
- Boeing wanted a rapid and fairly inexpensive development cycle; this is, after all a prototype.
- Computers need to be tested when you put them in aircraft, especially experimental ones where you have to throw out much of your previous learnings.
- In an experimental aircraft, the computer programming would have to change with each iteration of the aircraft; this 'tweaking' could cause bugs to creep in and would certainly add to the time required to create a new version of the craft.
Based on that, adding flight computers would have been expensive, time consuming and wouldn't add to the value of the experiments. It would also have added weight & power consumption to the craft, neither of which is desirable. If it were to go anywhere near production and real use, that is when you start looking at the computers.With the obvious problem that you're spending $4000 to back up a system which probably cost Given the cost of DVD burners + media, you're considerably cheaper looking at DVD for backup than any tape technology.
As for people saying "turn off the messenger service", there are actually valid uses for winpopups. At my last work, I set up a few perl scripts that would use smbclient to warn Samba users when they were over quota. Before that, users would go over quota and wouldn't know about it until things broke after the grace period.
Obviously, you should be filter Netbios ports at the firewall unless you have a damn good reason to have internet access to them. If someone in your network is using this program to spam, the LART them appropriately.