GNU Hurd Is Up To 344k Lines Of Code

[phoronix]

Phoronix: GNU Hurd Is Up To 344k Lines Of Code

While GNU Hurd has been around for years prior to the existence of the Linux kernel, Hurd is now up to only 344 thousand lines of code, but it's having a hard time getting much higher...

http://www.phoronix.com/vr.php?view=MTU1NTM

[You-]

Virtual OS

AT this moment in time I think its unlikely for this OS to become complete enough for daily use - Drivers and firmware are probably still a huge hurdle.

But those are not as necessary for a virtualised OS where it only needs good support for one set of virtual drivers (but will need to target the different VM technologies) and the host OS can take care of the rest. This way it can target eg servers.

[profoundWHALE]

AT this moment in time I think its unlikely for this OS to become complete enough for daily use - Drivers and firmware are probably still a huge hurdle.

But those are not as necessary for a virtualised OS where it only needs good support for one set of virtual drivers (but will need to target the different VM technologies) and the host OS can take care of the rest. This way it can target eg servers.

I think you're right, at least if we want development to speed up.

[randomizer]

Reformat the whole codebase to wrap all lines over 40 characters. Instant boost to the most important code metric.

[a2r-l]

AT this moment in time I think its unlikely for this OS to become complete enough for daily use - Drivers and firmware are probably still a huge hurdle.

I see what you did there

[TemplarGR]

Lines of code is the most overrated metric ever. I refuse to cooperate or take seriously ANYONE who thinks this metric is important...

[mrugiero]

Lines of code is the most overrated metric ever. I refuse to cooperate or take seriously ANYONE who thinks this metric is important...

If this is for randomizer's comment, I believe he is being sarcastic. I refuse to believe anyone considers there is a direct relation between good software and lines of code. IMO, for the same language and set of features (otherwise, comparing this is just 100% pointless), the more lines, the more mess.

[Delgarde]

But those are not as necessary for a virtualised OS where it only needs good support for one set of virtual drivers (but will need to target the different VM technologies) and the host OS can take care of the rest. This way it can target eg servers.

Still not good enough. We already have operating systems that do a perfectly good job of running servers in VMs, and which also do a fine job of running servers outside of VMs (among other things). What's the value in something so specialised it can *only* run inside a VM? That's a big loss of flexibility, for no obvious gain...

[philip550c]

Still not good enough. We already have operating systems that do a perfectly good job of running servers in VMs, and which also do a fine job of running servers outside of VMs (among other things). What's the value in something so specialised it can *only* run inside a VM? That's a big loss of flexibility, for no obvious gain...

Well there is this http://www.phoronix.com/scan.php?pag...tem&px=MTQ2NDI

[Delgarde]

If this is for randomizer's comment, I believe he is being sarcastic. I refuse to believe anyone considers there is a direct relation between good software and lines of code.

Oh, there's a relationship... but an inverse one. Sure, you can take that too far - the objective is clarity, not a reduction in newline characters - but as far as I'm concerned the shorter the better. By all means extract the details into sub-functions, but if the basic logic doesn't fit into one or two screen pages, chances are nobody understands it.

I maintain an awful lot of inherited code that fits that description, and inevitably, the first step towards making any changes to it (whether bugfixes or new features) is to start carefully simplifying until I can tell what it's doing... and whether that matches up with what it's supposed to be doing.