fighting the lack of good ideas

not enough cpus

I found an interesting reference that indicates that for x86, the maximum number of CPUs the Linux kernel can handle is 255. I’d presume that it’s the same for Windows.

I’m curious – right now this is a rather large limit: but it won’t be for long! 16 core CPUs are available from AMD today. I see motherboards that will support 4 sockets (there may be some that go higher). That’s a total of 64 CPUs visible to the OS.

With AMD and Intel pushing the number of cores ever higher, how soon will we hit the practical limits of the x86-64 architecture – or is the limit of 255 not a “core” limitation of the architecture?

I’m curious to find this out – and to see how vendors will handle these upcoming limitations.

symlinks and nfs

I recently discovered an interesting “feature” of symbolic links in conjunction with NFS mounts: they don’t work!

For example, let’s say you have the following NFS export:


Inside of that export, you have the following path:


In *this* directory, you have a symlink called ‘current‘, which points to a different location:

current -> /media/osmedia/linux/ubuntu/11.10

Here’s the rub: if anyone tries to mount the root (or sub path) of that NFS export (say /media/files/isos/osmedia) at a local point (say, /mnt/osmedia), the symlink will fail to work correctly because there is no path /media/osmedia/linux/ubuntu/11.10 on the local machine!

That’s not cool.

Here is another case where bind mounts will save your bacon!

Instead of using symlinks on the NFS host, use bind mounts – that way the path will still work when exported.

mounting a cd in an rhel6 vm

Most typically in the past, I have seen mounts for CDs in RHEL work thusly:

mount /dev/hdc /mnt

In RHEL6 you need to do the following:

mount /dev/sr0 /mnt

I’d really like to know why Red Hat made the change, but knowing is useful 🙂

vmware tools can be quietly installed

Came across this little gem recently:

/path/to/vmware-tools-distrib/ --default EULA_AGREED=yes

That’s been a life saver a lot recently 🙂

bagel [thin] burgers

Several years ago I had the good fortune to have a burger paired with cream cheese on a rye bun.

And a few years later I had a burger on a bagel. I’ve also enjoyed bagels on Arnold sandwich thins.

Combining the best of both of these ideas, try bagel burgers (or on bagel thins) – with cream cheese.

Grill your burger to desired doneness while lightly toasting your bagel [thin].

Apply cream cheese to bagel and add your freshly-grilled burger.


Of course, other favorite condiments and pairings could be added – bacon, mushrooms, jalapeño coins, etc.

creamy, cheesy baked potatoes

I don’t like sour cream (not when I know it’s there, at least).

However, I do like creamy baked potatoes.

And I like cheesy baked potatoes.

My solution?

Cream cheese instead of sour cream!


  • bake potato(es) to desired doneness
  • split in “Wendy’s” fashion (perforate with a fork, and squeeze the ends perpendicular to the perforations)
  • dollop whipped cream cheese into opening
  • mix in
  • add bacon bits, other cheese, chives, etc to taste

The cream cheese also does a good job of substituting for any butter you might otherwise have planned to add to the potato.

Be sure the cream cheese is of the whipped variety – solid cream cheese takes too long to mix in 🙂

fixing copyrights and patents

Following-up a recent post on copyrights, I want to share some further thoughts I have on the topic, and about patents, too.

First of all, the concept of a copyright is meant to protect the author from others unduly benefiting from their work. One obvious conclusion to make from that statement is that after the author ceases to live, they can no longer claim to receive any benefit from their work. I know I certainly wouldn’t care about royalties after I die.

Second, only individuals or their proxies should be allowed to claim copyright over a work. By proxy, I would include work created for a company in the context of something sold and/or shared publicly.

My proposed fix for copyright law would be to cap copyrights at 25 years, or the life of the author – whichever is shorter. If something is written for a corporation, the copyright could be retained either by the author(s) or the company, but it will expire no more than 25 years after it has been written (this would cover the case of one of the authors being tragically lost due to illness or accident). Knowledge grows when information is spread. The more people have access to information, the more applications of it can be made. And, overall, knowledge is a Good Thing™!

With regards to patents, I think there are many broken aspects of the current system: not the least of which is that non-physical “things” can be patented (algorithms, software, etc). While not an inherently bad thing to be able to protect a proprietary process or method for accomplishing work, the fact of the matter with regards to modern society (speed/quality of communication, the ability to analyze data, etc) is that whether you patent your process for accepting multiple inputs at once to a program or not, someone will be able to [nearly] instantly copy what you did.

My first fix to the patent system would be to cap patent life at 10 years from the date of issuance, and 15 years from the date of filing (honestly, if it take more than 5 years to get it issued, your competitors have already caught and exceeded you). I would also preclude the ability to extend the life of a patent through any means – innovate more if you want, but you cannot extend the life of a patent past its expiration.

My second fix will be to ban the ability to patent software. It would not be just to apply a new law to old thinking, so extant software patents would not be affected beyond the ability to extend their life.

My third fix would be to ban security orders being placed on a patent (as the NSA has been known to do with regards to encryption algorithms). No one – private individual, corporation, or government agency – should be allowed to “preview” patent applications and attempt to get the blocked or hidden.

My final fix to patents would come with streamlining the application and approve-or-deny process. The USPTO is overwhelmed with applications. Some of this comes from companies trying to file exceptionally-broad applications just to see what they can get away with (“I know! We’ll patent the process of processing patents!” – or some other silliness). Some of this comes from inadequately-informed patent officers – there is no possible way every patent officer can possibly know about all the fields that patents are asked-for and -about! There needs to be an improvement in the general population of the patent office, whereby more skilled/knowledgeable/talented people are put in place to review patents (not saying they don’t have talented people now, but that needs to be increased).

Lastly – and this would related to copyrights, trademarks, and patents – a public database of all current patents and trademarks should be made available. That database should also show all expired patents and trademarks. And, for those authors who have chosen to register their copyright with the USPTO, a database of authors, their work(s), and the copyright date (and, by calculation, its expiry) should be visible.