[HN Gopher] Stone knives and bear skins - there is no money in t...
       ___________________________________________________________________
        
       Stone knives and bear skins - there is no money in tools
        
       Author : sizzle
       Score  : 25 points
       Date   : 2023-07-25 21:18 UTC (1 hours ago)
        
 (HTM) web link (queue.acm.org)
 (TXT) w3m dump (queue.acm.org)
        
       | sanxiyn wrote:
       | Being against a kernel debugger is one of the worst decision
       | Linux ever made. Thankfully those days are behind us.
       | 
       | https://lwn.net/Articles/270089/
       | 
       | Linus Torvalds, in paraphrase:
       | 
       | > I happen to believe that stone knives and bear skins force
       | people to think about their problem on a different level.
       | 
       | This is exactly as ridiculous as it sounds.
        
         | kbenson wrote:
         | It's ridiculous, but at the simplest level, it's not wrong.
         | Your tools or the lack thereof will change how you think about
         | things, for better _or worse_ depending on the situation.
         | 
         | If your goal is putting venison on the table tonight, vs
         | understanding deer behavior and herd migration and when deer
         | are even in the area in enough numbers and so can even be
         | thought of as a viable food source, a stone knife and a bear
         | skin as tools might promote (or force through necessity) better
         | understanding than a hunting rifle and a large puffy jacket.
         | 
         | That doesn't mean it's "better" or that everyone should be
         | using shitty tools, but it is food for thought, and that
         | perhaps doing it a few times is a learning experience that
         | isn't all negative.
        
         | m463 wrote:
         | kernel debugging on linux sucks (IMHO)
         | 
         | I've tried it and it is really hard to get even the most basic
         | debug output - a symbol'd stack trace of a crash.
         | 
         | If you can't get it at runtime, you need a kernel dump, and
         | trying to get one is a mess. Once you have it, it is hard to
         | match up the dump to the symbols. I believe some of the tools
         | were kernel-version specific and I think when kernel 5 came out
         | they didn't work.
        
         | JohnFen wrote:
         | Yeah, I think he overstates the case. I do see people who have
         | become overly dependent on debuggers and that has adversely
         | affected their debugging skills -- but that isn't an indictment
         | of debuggers.
        
       | version_five wrote:
       | > A venture capitalist once told me, "There is no money in
       | tools."
       | 
       | That suggests he's saying there is not upside in a portfolio of
       | tool companies. Following average VC advice gives average VC
       | results. If everyone listened to this kind of stuff we'd be awash
       | is generic SaaS CRM stuff...oh
        
       | honeybadger1 wrote:
       | While this essay serves as an interesting critique of the current
       | state of operating systems and their development, it perhaps
       | falls prey to the very flaw it seeks to highlight - the tendency
       | to oversimplify the complexities inherent in systems software
       | development and to yearn for a 'perfect' solution that is largely
       | a figment of the author's imagination.
        
       | ggm wrote:
       | There no money in tools: The Dutch company ASML would like to
       | talk to you.
       | 
       | The article does actually say that in highly specialised niches
       | tool making is fantastically lucrative. So, don't mistake the
       | headline for ground truth.
       | 
       | Also there is good reason to believe tool making since flint
       | knapping has been high status, and demands surplus production to
       | liberate skilled labour to making tools which have immensely high
       | use value: tools are magnifiers of human productivity and are
       | worth a lot.
       | 
       | The point is that when commodities are cheap, and tools like gdb
       | can be considered a commodity, the relative advantage of "a
       | better one" has to be weighed up with using a screwdriver as a
       | hammer. Thus printf() debugging.
        
         | trollerator23 wrote:
         | In _software_ tools. Hardware tools are hardware.
        
         | JohnFen wrote:
         | Well said. I know there's money in tools because I've been
         | making money making and selling tools for a long time.
         | 
         | But not every kind of tool. A gdb replacement would be a hard
         | sell unless it managed to actually perform miracles. But
         | specialty tools of the sort that can save a relatively small
         | group of devs major pain and/or time? There's good money in
         | that for a small business. That the market is too small to be
         | attractive to large companies is just gravy.
        
           | sanxiyn wrote:
           | Time-travel seems miraculous to me, but uptake seems very
           | slow to me.
           | 
           | I hope https://www.replay.io/ succeeds.
        
             | Veserv wrote:
             | Commercially available time travel debugging is over 20
             | years old at this point [1]. So yeah, the uptake is pretty
             | slow.
             | 
             | [1] http://jakob.engbloms.se/archives/1564
        
         | 36870059 wrote:
         | ASML's crown jewels are not mere "tools", they are _machines_.
         | It is bleeding edge _technology_ and entirely irrelevant to a
         | general discussion of economics of  'tool makers'.
         | 
         | https://www.youtube.com/watch?v=5Ge2RcvDlgw
        
           | ggm wrote:
           | and GDB isn't a machine?
        
             | 36870059 wrote:
             | Anyone can make a GDB. Haven't you noticed? _Superpowers_
             | are fighting over ASML..
        
       | m463 wrote:
       | I think there are a few things at work.
       | 
       | There is no money in tools because employers (and to some extent
       | employees) don't value them. Also, folks who make tools maybe
       | value them too much? And maybe tools are too easy to copy?
       | 
       | I remember decades ago when purify came out. I tried the demo at
       | work on a project I was working on. Within a few hours, I found
       | an enormous number memory errors. It was enormously helpful.
       | Purify had a high per-seat license cost, and work wouldn't buy
       | it.
       | 
       | Many current employers are a little smarter about things, but I
       | mainly see things like code coverage tools and version control
       | systems, but no real spending on more personal/fashionable tools,
       | even if employees are more efficient.
       | 
       | I think tools that might work better would be those employees
       | bought them themselves - think jetbrains - and priced
       | accordingly. The analogy might be auto mechanics building up
       | their collection from snap-on.
       | 
       | also, warts and all, linux is a nice environment because you can
       | get to the bottom of most things and tools might be less
       | industrial quality, but there are plenty of them.
        
         | neilv wrote:
         | Purify was glorious. As was Saber-C before it.
         | 
         | My employer paid for both. But we sold even more expensive
         | development software ourselves, and bought very expensive
         | workstation and server hardware to run it on, so we were
         | accustomed to the price points.
         | 
         | I like that great software is now open source, and most any kid
         | or adult around the world can run it on a cheap decade-old PC.
         | But I do miss some things about that software development era.
        
       ___________________________________________________________________
       (page generated 2023-07-25 23:00 UTC)