Posts by guyr 83 publicly visible posts • joined 25 Oct 2012
Like others, I use VirtualBox for my personal VM needs, have been doing so for over a decade. So, I don't really understand why free VMWare is such a big deal. I guess if you use the paid version in a work environment, having the same available as a free option for non-work environments makes sense. Would have made more sense 10 years ago before people were forced to find other options.
Doug 3: "If an application wasn't multi-threaded(all OS/2 apps were required to be ) you could switch to another app while that one finished doing its work."
Not true. I worked on an IBM OS/2 app from 1991-96. I was in Hong Kong on a customer support assignment. Scanner code was locking up. By running the code in a debugger, we were able (with some remote support from a large company in Redmond) to determine that the scanner code had been linked incorrectly, without the /MT flag. So, OS/2 apps were not required to be multi-threaded, though the overhead was so slight, that *should* have been the default behavior.
The task switching you describe worked because the OS itself was fully multi-threaded. Though even to the very end, OS/2 (including 2.0) had a single input queue for desktop GUI events - mouse clicks, etc. - so completely locking up the desktop was very easy and happened frequently. That was also true of Windows, but MIcrosoft finally fixed it in NT.
I also got most of my formative Linux graphical experience with GNOME 2, so when I saw the proposals for GNOME 3, I investigated how I might stay with GNOME 2 without excessive work on my part. MATE fit the bill for me. While MATE has since migrated to GTK 3+, it continues to provide the traditional look and feel of GNOME 2. As an added bonus, it is lighter weight than many other UIs, so runs well on (slightly) older computers.
I worked most of my career as a software developer in a corporate environment, i.e, software for internal use. In 35 years, the top priority was always "get it done". Perhaps Intel is only targeting its efforts at huge software vendors like NVidia or the game companies. But for corporate developers, these types of specialized routines will not be used unless they get baked into the common development tools used: Java and C++ tool chains, etc. If we can get a 25% speedup (or a 25% reduction in resource use) by changing a compiler flag, that will help. If we *might* see an improvement by manually stitching an Intel-supplied module into our software build (which we then have to test extensively to make sure nothing else breaks), that will never happen.
Wasn't that the selling point of v6? If we have to randomize our v6 IP address on a daily/hourly basis to avoid tracking, doesn't that greatly diminish the value of having such an expansive address space? I guess if we can hide all that inside the ISP connection point (normally, an ISP-provided router), then we can just use NAT on all home devices and be done with it. As long as we can continue to use v4 inside the house for ease of use, and NAT that to v6 on the ISP router, that should cover the majority of use cases for home users.
Professionals get paid to deal with these headaches, so I'm not too concerned about the hoops that they have to jump through with all this.
Hmmm, how about actually making chips in the US? Or perhaps the US Department of Commerce is unaware that over the last 2 decades, just about all state-of-the-art production has migrated to the Far East? I'm surprised that we have constrained (or no?) domestic capacity for 40nm to 250nm production. The US had lots of equipment with that technology 20-30 years ago. But I guess fabs can't just sit idle waiting for customers to show up.
Rust is not trying too hard. Rust was designed to be a systems-level language that addresses well-known weaknesses in other languages. You can read widely available resources on the whole breadth of Rust design goals, but the primary one everyone talks about is memory safety. Operating systems have been battling memory access issues forever; Windows is coming up on 40 years old, and still struggles with hacks achieved through buffer overflows, etc. So, if you could eliminate this entire category of vulnerabilities by the design of the programming language itself, that would be huge.
You are correct that you can't correct program logic errors in a programming language. But Rust helps address other major problems like the memory safety issue.
Hmm, a modern CPU can have anywhere from 10-35 billion transistors, and take 3-4 months to manufacture starting from a polished wafer. I'm having trouble envisioning any CPU having significant portions of these precious transistors sitting non-functional in the anticipation that a customer *might* decide somewhere down the line to enable additional features. *Maybe* take a gamble with 10% of available transistors. But with global production constrained and customers lined up for products, the vendor (AMD, Intel, etc.) would be much better off using those idle transistors to make additional CPUs.
But the people running these businesses are undoubtedly much smarter than I am, and likely have decades of accumulated data showing which features customers are likely to want.
I actually am getting old, but I can't get excited about either Windows 11 or the next mobile standard 5G. My main computer is still running Windows 7 (I have a secondary on Windows 10, but rarely use it), and my 3 year old 4G phone on Android 10 does everything I require. Microsoft positioning Windows 11 not to run on anything older than 2 years makes my (non)-decision about Windows 11 very easy.
Well, it's very convenient for developers running in a corporate network that's using Active Directory and Windows clients. You can have several entire Linux distros of one's choice on that client, and you need never bother IT for permission to put a Linux machine on the corporate network.
Been doing that for well over a decade using Virtual Box. And VB has supported GUI apps - including seamless mode which displays Linux windows independently directly on the Windows desktop - for many years. So, what unique capability does WSL add?
TSMC has stated they are spending $20 billion on their next generation fab. I would imagine that is a significant hurdle for U.S. companies. TSMC gets all sorts of subsidies, which the US doesn't do. I think this will require some law changes as well, to allow US companies to cooperate without fear of all being sued for anti-competitive practices. I know Intel has said they are looking to build two fabs in Arizona; I don't believe it, I think they are just trying to exert economic pressure on TSMC and others. Intel has already said their lower end processors are being farmed out to TSMC. These fabs are just too expensive for individual US companies to undertake without subsidies or other financial support.
I thought HBM was designed to address this problem. You'll never be able to solve this problem generally via merging memory into the CPU, because different applications need different amounts of memory. The latest Intel Xeon Cascade Lake has "only" 38.5 MB of onboard cache. So, figuring out how to expand HBM seems to be a better expenditure of time, energy and money.
init option in sixth birthday release
If you have Internet, you have fax. :) I've been using fax1.com for years because there is no set monthly cost, and when you buy credits for faxing at 12 cents (US) /page, it never expires. Would likely get expensive if you fax in high volume, but I bought $10 probably 8 years ago and haven't used it up yet.
If you have Internet, you have fax. :) I've been using fax1.com for years because there is no set monthly cost, and when you buy credits for faxing at 12 cents (US) /page, it never expires. Would likely get expensive if you fax in high volume, but I bought $10 probably 8 years ago and haven't used it up yet.
This is the point people need to acknowledge. All the major semiconductor companies - IBM, Intel, AMD, etc - did their manufacturing in the US.All those plants actually still exist, and are churning out chips today, but not the latest and greatest. Why? State-of-the-art chip manufacturing is not for the faint of heart. TSMC's current 7 nm fab cost $10 B, and the next one is estimated to cost $20 B. For comparison, AMD had a great year in 2020, with a net income of $2.4B, but that included a one-time income tax benefit of $1.3B, so more like $1.1B, which is still great for AMD, which had lost money for 15 years straight.
So, how does a company who makes $1.1B in their best year afford a $20B manufacturing plant every 10 years? It can't, math doesn't work out. AMD outsourced manufacturing to TSMC for its latest chips because it had no other choice, not because it just wanted to make a couple more bucks. Building those chips on their existing (last-generation ) production line would have been unsuccessful - too big, too much heat. And they didn't have the time or money to build a new current-gen fab.
So, how does TSMC do it? Government subsidies and indirect government support (tax breaks, low interest loans, etc.) TSMC can easily get that $20B loan and go 10 years showing little or no profit. No US company can do that. So how do we fix this? I'm not an economist or a miracle worker, but I doubt we can mass-hypnotize American investors to accept zero profit in their investments for 10 years at a time. So, off the top of my head, I'd say we need to reorganize how massive investments can be made while still showing profits. Public funding? Accelerated tax write-offs? Smarter minds than mine need to think about that.
Despite Microsoft pouring billions into loss-making media ventures, creating new options for advertisers inside Xbox games and boasting annual search-driven ad revenue in the billions, Smith blamed others for the media’s problems.
None of which has anything to do with news. I'm not a shrill for Microsoft, Google or Facebook, but CNet could help understanding by focusing on the topic at hand.
"My guess as to what's gone wrong isn't 'malloc' as such but 'new'. This can create quite large objects either from pool memory or on the stack"
"new" does not allocate objects on the stack. But the larger point is that while I haven't looked at the httpd code recently, most code written in the last 10-15 years doesn't use "new" directly, but instead relies on libraries to manage dynamically allocated memory. The problem of memory not being freed has been well-known for decades, so projects of any significant size use libraries to ensure that doesn't happen.
At least Mate (based on GTK 2 last I checked)
Mate moved to GTK 3 several versions ago. I use it exclusively on all my Linux systems. It's original design goal directly from the MATE web site https://mate-desktop.org/:
"The MATE Desktop Environment is the continuation of GNOME 2."
So, the look and feel of GNOME 2 is the essential ingredient, rather than a particular generation of GTK.
I've been running Ubuntu MATE (after starting with Mint MATE) for several years, and it has been trouble-free. You can opt for the LTS (Long Term Support) releases, which only come out every 2 years, and thus provide a more stable platform. In reality, I've found that MATE provides its own stability, so don't find the 6-month releases that disruptive. But I've adopted LTS on all my systems except for one VM I use expressly to see what's currently going on, simply to cut down on the update volume.
So the last generation of desktop processors were the 3000 series, why has AMD jumped to 5000? I see on this page:
https://www.amd.com/en/products/ryzen-processors-laptop
AMD appears to have allocated the 4000 series to mobile chips. Is that a one-time thing, or are they going to be allocate even-numbered series to mobile and odd-numbered series to desktops?
And then there is this head-scratcher. On that page: "AMD Ryzen™ 3000 Series Mobile Processors with Radeon™ Graphics", under which is ... wait for it ...
AMD Ryzen™ 9 4900H, AMD Ryzen™ 7 4800H, etc. How's that for clear marketing???
Took a look at the specs - missing bands 66 and 71. So, forget about 5g. Unfortunately, this phone doesn't even have newer 4G bands. I like the idea, and my needs are pretty basic; I especially like the replaceable battery, which every phone before my current one had. My current phone is LG G7, which was engineered in 2018, and has both 66 and 71. So, doesn't seem like including these bands in 2020 should be a difficult engineering challenge.
Java is cross-platform, and is by far the largest language used for enterprise applications because it can get you were you need to be in a reasonable amount of time and resources. But Java GUI apps never really caught on because, as you say, they tend to look terrible and non-native, so people don't like them.
If Flutter can get an app on desired platforms in a short amount of time to get a feel for potential user interest, that seems like a desirable approach. Most organizations either can't afford or don't have the time to simultaneously develop for multiple native platforms. Then if sufficient interest exists, developers can weigh the cost and benefits of a native approach for specific platforms.
"AMD is also playing nice with Oracle, with the chip slinger's Epyc Milan silicon on the way. Oracle plans to rent individual cores of the new processors to those running microservices and similar workloads that can thrive on modest resources."
This is an odd statement. AMD is presently outperforming all available Intel options by a wide margin. So why is Oracle targeting AMD cores only to workloads requiring modest resources?
I worked in software development for 40 years, now mostly retired but dabble in open source projects to keep my brain working. In the early years, we used rudimentary editors and makefiles. Got the job done, but the toolset was not really helpful to comprehension.
For the last 25+ years, we've used IDEs for the development task, which greatly facilitates understanding. Tools don't replace the need for a developer to understand what the code is doing, or even more important, fundamental software concepts. But it certainly helps competent developers to do their jobs with fully cross-referenced source, source-level debuggers, etc.
To your point, plain text facilitates accurate communication between remote systems. Today we have tremendous tooling that enables easier communication between people. Think distributed reviews.
Ye of little faith. Please see my comment on the page in the following link. Hyper-V can be enabled on Home with some easy to follow steps.
https://docs.microsoft.com/en-us/answers/questions/29175/installation-of-hyper-v-on-windows-10-home.html
Yes, that's it exactly. I found this topic interesting enough to do some searching, and found this article that explains some of it. Apparently that band is used by wireless microphones as well as microwave backhaul links. To us not in the industry, WiFi in the home wouldn't appear to impact either of those applications. However, I can see a scenario where someone is trying to do an on-site news report in a suburban neighborhood, and every house is leaking these WiFi signals, rendering the spectrum unusable for licensed users.
https://www.sportsvideo.org/2020/05/12/spectrum-faces-its-next-challenge-as-fcc-allocates-6-ghz-range-for-wi-fi-6/
Jim Keller is without doubt a very smart man, given his stints at every major CPU manufacturer. But he doesn't appear to like to stay in one place for very long. Given how complex the task is, I'm a little surprised the big names - Intel, AMD - are willing to make the investment. Keller joined Intel in April 2018, meaning he's been there a little over 2 years. How quickly can anyone absorb an architecture with billions of transistors to start making significant contributions? I worked in software development, not hardware, and we would allow new people a couple months to get familiar with the architecture of any project before we expected major independent work.
>> "Meh. Pascal appealed because it was strongly typed."
Indeed, others here are criticizing Pascal being strongly typed. They fail to recognize that this one feature is what made Pascal the preferred teaching language in the 70s. I went to a pretty rigorous engineering university, and they employed Pascal in both intro and advanced language classes. If you are instructing people who know nothing about structured programming, then you want to start them thinking in very methodical techniques. Pascal provided that structure then.
In my experience, the responsibility goes further than that, to the carrier. In my experience, for an Android update to appear on my phone requires the cooperation of all 3: Google, the phone manufacturer, and the carrier (T-Mobile in my case.) Unless all 3 are on board, the update will not show up on the phone.
I came here to say the same thing. We may *want* to get updates, but unless the maker of our phone and our carrier both provide the updates, then it ain't happening. Telling phone users to get an update accomplishes about as much as telling your dog to get a driver's license.
"In LibreOffice, you can change chart type and hide or show the legend and that's just about it."
I've only used LO at home, where my needs are not that great. But I did add a chart to my retirement spreadsheet, and tweaked to get what I wanted. The chart capability actually does provide quite a bit of customization, though it might not be as easy to find and apply as you'd wish. As one example I remember off the top of my head, to tailor how individual lines in a data series appear, you can't just right click the line and modify properties. Instead, you need to open the data series editor and change properties there.
To think that we're going to be forever lumbered with this sad example of "eh, good enough" design saddens me when there are so many better systems that already exist.
Forever? Not if somebody thinks of something better. Remember, Torvalds developed git as a 1-month side project when he couldn't find an available version control system that worked as he liked, without unbearable licensing terms. The entire open-source universe has proven its meritocracy-based roots. Torvalds himself would not hesitate to throw out git if something provably better comes along.
"This article is exactly about now being able to develop GUI application for Linux and no longer only console ones."
Thanks for the correction. I went back and reread this part:
"FmxLinux was developed by a third party, Eugene Kryukov. It has been licensed under “a long term distribution agreement,” says Embarcadero’s Marco Cantu in the announcement this week." and followed the link to the announcement Makes more sense now. But I also read this on the announcement page: "active subscription to the Enterprise or Architect editions". What's up with Embarcadero? They really like to put the screws to their customers. Less pain-seeking customers can just license FmxLinux directly at a much lower total cost.
"On the Ubuntu side, you have to install an agent which communicates with Delphi on Windows."
What's this? I took a 15-minute look at Kylix when it came out, and honestly I don't remember anything about it. But this quote has me scratching my head. Am I understanding correctly that to run a Delphi app on Linux, you need to shuttle work over an agent to a process running on Windows? Sounds like the mother of all kludges. I really can't see the niche for this. If you have a staff of experienced Delphi developers, and you want to deploy a bespoke app onto a Linux desktop? But if you also need to have a full Windows running to do the grunt of the work, why bother? Just run the app on that installation of Windows.
Okay, found this article on the Embarcadero site that explains things:
http://docwiki.embarcadero.com/RADStudio/Rio/en/Linux_Application_Development
The agent is only required during development, not deployment. You develop the app on Windows, and cross-compile it for Linux. GUI apps are not supported, only console apps. Meh, very niche market. If you want to run a full app on both Linux and Windows from a single source, several options are available. Java of course, but for native look and feel, WxWidgets gets you there.
I was inside the castle for pretty much the entire run of OS/2, from pre-1.0 to 3.0 Warp. From my perspective, what killed OS/2 was IBM's defiant adherence to the original design decisions, and a refusal to modernize. For example, IBM insisted it run on the 286, even though by the time it reached any measurable market, the 386 was ubiquitous. That let to numerous architectural restrictions the messed up the works. From a developer's viewpoint, the single input queue (from all Windows versions prior to NT) was a nightmare; this was the old cooperative multitasking model, where a single badly designed program could hang the entire OS.
Microsoft addressed all these issues in NT, but IBM refused to budge with OS/2. When Microsoft demonstrated they were serious about NT, everyone gave up on OS/2. It held onto that darned single input queue to the very end!
I doubt mainstream businesses upgrade CPUs very often. If a company has a rack of dual CPU servers supporting their business and they are working acceptably, they don't risk disrupting their business by tearing apart that rack. Instead, they'll just keep those servers going until they are depreciated, then replace them completely with newer ones.
The only large scale upgrades I've ever read about were supercomputers, where they bring the system offline and replace (e.g.) 8096 processors at once. They can do that because time on that system is scheduled, and it's not used for daily purposes like processing credit card transactions.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
