Category Archives: fedora

all about fedora

Simple life hack

Turn your mobile data to the OFF as a default when out and about.

Why? Read on.

Since about October 2017, I’ve been quite conscious about how my phone connects to the data network via the mobile provider. There was a time I did not bother about it and having the instant connection to the Internet that I love and can’t live without, came at a price. The price of being constantly on and constantly distracted. It was getting to be too much. We see this everywhere.

I had to do something, if anything, for myself. My phone’s mobile data is now set to default OFF when I am out and about.

The advantages are:

  1. You get back your time. You can then enjoy the world around you.
  2. You have control as to when you can be contacted. My family knows that I will turn off the data when I am outside and turn it on only when needed. If there is anything they need to contact me on, they can first SMS (remember that feature?) and then I can turn on the data and do what is needed.
  3. Your mobile data usage is not going to surprise you by being excessive. In face, my average for the last 1 year and a bit of doing this, has been just about 1GB.

Remember, there is nothing so important that you must have your mobile turned on at all times. If you can’t be reached, there are other means if it really matters.

I would be remiss if I don’t also mention that turning off the mobile data as a default is also driven by the fact that there are now plenty of wifi hotspots that one can access. In my morning commute from home to the office which takes about 35 minutes or so by bus, I come across two spots where Wireless@SGx is active. As the bus enters those zones, viola, my phone connects automatically (yes, Wireless@SGx works flawlessly for me on my Android and Fedora systems). If the bus was around these hotspots for long enough, I get messages coming in and I can choose to do what I want. Most times, it is those that are quite superfluous and would have distracted me otherwise.

When I travel, I do not subscribe to the expensive data roaming plans that Singttel offers. They have a stupid way to compute and I do two things: a) use wifi where available and b) get a local SIM (data only if possible, else the usual types). Fortunately, I do not have any need to be reached on my Singapore phone number and if needed, I can be reached by other means first and I can then get on to a voice call using many of the tools out here like Jitsi, or Signal, or Telegram. I have VOIP set up as well via LinPhone on my mobile and laptop, so I am pretty much covered if I really need to do a voice chat.


Not allowed to code? Really?

[image from:]

Lots of interesting, but not surprising, information is being made public about the Singhealth data breach.

The Commitee of Inquiry has been told that there was an IHIS employee who found a bug in the Allscripts “Sunrise Clinic Manager” EMR in 2014 who then made the loophole known to a rival of Allscripts, Epic Systems Corporation. Both of these vendors products are closed, proprietary and, IMHO, unnecessarily and excessively expensive products.

From this report (again, this is MSM reporting, so take it with two pinches of salt – you have to read the court transcript which I am not sure if available, yet, if ever), says that the IHIS employee was “unhappy” that he could not do coding in the job role he was doing and so, he then decided to contact Epic to tell them of the issue so that it could “… leverage the vulnerability to gain a larger market share” (emphasis mine).

Larger market share? How so? The hospital clusters in Singapore are about evenly locked-in between these two proprietary vendors. Moving from one to another is not a simple thing. And one bug does not even start the thinking process. Who knows if the Epic product has similar issues?

Not being able to ascertain if the reason offered by the former IHIS employee is indeed valid, I find that it seems to be a fluffy afterthought. Having been caught out, the former IHIS employee is offering excuses.

Not Allowed To Code?

I find that reason to be intriguing. Did the job that the IHIS exmployee took on involve coding? No indication in the report. If that was what the person wanted to do, why not channel the skills an open source project that could use help? No one will stop you from doing that, unless, the terms of employment of IHIS says that a developer “cannot work on any software project other than what is part of the job”.

I have no insights on what the terms of employment are, but here is an example of an enlightened and correct way to encourage developers:

“Participation in an open source project, whether maintained by the Company or by another commercial or non-commercial entity or organization, does not constitute a conflict of interest even where such participant makes a determination in the interest of the project that is adverse to the Company’s interests.”

– taken from page 3 of

Software developers are artists. Software development is an art form. One would not constraint a painter, so why would one shackle a software developer?

Bug Reporting, Fixing and Regression Testing

If a bug is reported – whether it is a “the button is of the wrong shape” or “this option dumps out the entire database”, assuming that proprietary vendors have a bug reporting process – nope, they don’t – then things can be moved along without too much excitement. All software have bugs. If a vendor (open or closed) does not offer a way to report bugs, you have to demand that there is a way to do it.  Red Hat has both and to submit bug reports on all of the open source projects and open source products (go here for an understanding of the differences between open source projects and an open source products) that Red Hat is involved in and makes available to paying customers (

Maybe there is a some place at Allscripts and at Epic Systems that one can file bug reports, but it is not immediately evident.

Regardless of being able to report bugs, I do wonder how these vendor organizations manage bug reporting/fixing and regression testing. I have to assume that they do it properly (for some definition of properly) but it is telling that a trainer of Allscripts said this:

“Another witness, however, called the loophole “perfectly normal”. Mr Loo Yew Tuck, senior lead analyst at IHiS’ clinical care department, said that he had seen an Allscripts trainer demonstrate its use and method previously.”

Really? There is a “perfectly normal” loophole? Or did he mean, backdoor (of the NSA type)?

I particularly concerned with this paragraph – as reported in another MSM report

“… She also did not know the details of the alleged loophole. Neither did she ask her staff for it to be verified. She also assumed that the problem would be rendered “irrelevant” as IHIS had just upgraded the EMR system architecture”.

If the bug is not reported, how would one know if it was really an issue and if so, if it was indeed fixed? Granted, we cannot all be on top of things all the time, but if there isn’t a process to track issues, what then?

“… did only what … was asked to …”

Leadership and empowerment failure. Whether it is real or otherwise it is hard to tell. Perhaps there is a culture of empowerment but not everyone got the memo. Of maybe not. I can’t tell.

Happy 35th Birthday GNU!

The GNU project was officially announced on 27 September 1983 by Richard Stallman. Thirty-five years of a project that has now become the fundamental building block of everything we use and see in technology in 2018. I would not be wrong to say that there isn’t a single proprietary piece of software that anyone is still using from 35 years ago – please post comments if there is something still being used.

There is only one reason for this longevity: the GNU project was built upon the premise that the code is available to anyone, anywhere with the only restriction that whatever is done to the code, it shall always be available to anyone, forever. Richard Stallman’s genius in crafting the copyleft license is the GNU General Public License is probably the best hack of the 20th century software industry.

What was I doing in 1983? I was working at the Computer Systems Advisors (no longer around, after being bought up by CSC). I was working on a system from DEC – the PDP-11/44 running RSX-11. We were creating software using Digital’s DIBOL – a variant of COBOL. DIBOL was, IMHO, the sink-hole of GOTO statements. I truly did not like that language. I can’t pinpoint why, but it just did not appeal to me. I had learned FORTRAN and BASIC before and they were so much more expressive. Editing the code in the PDP-11 was using some editing tool (perhaps teco or something) on a VT-100 terminal that was attached to the RSX-11 system. I think I was working on some insurance company’s bespoke code – but, frankly, those are all wishy-washy to me now.

The IBM PC was becoming an interesting offering and CSA’s sister organization, Automated Systems, got the distributorship of the Corona Data Systems PC. I got asked to take on the Corona PC as a main product person, I began my involvement with the PC-DOS/MS-DOS/CP/M world. I was learning a lot about this world with the ISA cards, Interrupts, RAM cards, RS-232 terminal connectivity, writing Interrupt Service Routines, terminate-and-stay-routines, and later 3Com‘s network cards – 3C501, 3C509 etc) with coaxial cables (yes, you better terminate the end of the bus with a terminator plug or else the network freezes). I recall that there was some networking software to share files (perhaps 3+Share or the like).

In parallel, the reason for the PC (including the Apple ][) getting traction in the early 1980s, was the famous spreadsheet product called VisiCalc created by VisiCorp. Their success encouraged VisiCorp to create something else called Visi On. This was an interesting product that provided a GUI for the IBM PC.

Automated Systems managed to get the distributorship of Visi On and I was then tasked with understanding the product and helping to get the training/sales going. There were many things I liked about it, but there was a lot more I could not understand because of the lack of documentation – or rather, insufficient documentation – on how to program in it etc. I recall a training session for a customer who specifically asked about creating macros for Visi On like one could with VisiCalc. To whoever that asked me then, I still don’t know.

I am sad to note that none of those closed, proprietary products are around. The loss of all of the good work done by the developers of those products is disappointing. If they had the foresight of Richard Stallman and made the code available, we might be on a different trajectory with technology than where we are now.


A healthcare IT foundation built on gooey clay

Today, there was a report from the Solicitor General of Singapore about the data breach of the SingHealth systems that happened in July.

These systems have been in place for many years. They are almost exclusively running Microsoft Windows along with a mix of other proprietary software including Citrix and Allscript.  The article referred to above failed to highlight that the compromised “end-user workstation” was a Windows machine. That is the very crucial information that always gets left out in all of these reports of breaches.

I have had the privilege of being part of an IT advisory committee for a local hospital since about 2004 (that committee has disbanded a couple of years ago, btw).

Every year, budgetary proposals for updates, new versions etc., of the software that the advisory committee gets for consideration and possible approval. Almost always, I would be the exception in the committee in questioning the continued use of expensive proprietary software for these healthcare systems (a contributory factor to increasing health care costs). But because I am the lone contrarian voice, inevitably, the vote will be made to approve and hence continue, IMHO, the wasteful path of spending enormous amounts of monies in these proprietary systems.

I did try, many times, to propose using open source solutions like, for example, virtualization from KVM. This is already built in into the Linux kernel that you can get full commercial support from Red Hat (disclosure: I work for Red Hat). You pay a subscription and we make sure that the systems are running securely (via SELinux for a start) and that enterprise can focus on their core business. But no, they continued with VMware.

I did propose open source solutions like OpenEMR and many other very viable solutions for the National Electronic Medical Records system – but none of them were accepted. (It has been brought to my attention that there are plans to mandate private sector healthcare providers to use the NEHR. There is considerable opposition to it both from the hassle (from their point of view) and added costs since the solution is proprietary and expensive).

There were some glimpses of hope in the early years of being on the committee, but it was quickly snuffed out because the “powers that be” did not think open source solutions would be “good enough”. And open source solutions are still not accepted as part of the healthcare IT architecture.

Why would that be the case?

Part of the reason is because decision makers (then and now) only have experience in dealing with proprietary vendor solutions. Some of it might be the only ones available and the open source world has not created equivalent or better offerings. But where there are possibly good enough or even superior open source offerings, they would never be considered – “Rather go with the devil I know, than the devil I don’t know. After all, this is only a job. When I leave, it is someone else’s problem.” (Yeah, I am paraphrasing many conversations and not only from the healthcare sector).

I recall a project that I was involved with – before being a Red Hatter – to create a solution to create a “computer on wheels” solution to help with blood collection. As part of that solution, there was a need to check the particulars of the patient who the nurse was taking samples from. That patient info was stored on some admission system that did not provide a means for remote, API-based query. The vendor of that system wanted tens of thousands of dollars to just allow the query to happen. Daylight robbery. I worked around it – did screen scrapping to extract the relevant information.

Healthcare IT providers look at healthcare systems as a cashcow and want to milk it to the fullest extent possible (the end consumer bears the cost in the end).

Add that to the dearth of technical IT skills supporting the healthcare providers, you quickly fall into that vendor lock-in scenario where the healthcare systems are at the total mercy of the proprietary vendors.

Singapore is not unique at all. This is a global problem.

Singapore, however, has the potential to break out of this dismal state if only there is both technical, management and political leadership in the healthcare system. The type of leadership that would want to actively pursue by all means possible to make healthcare IT as low cost and yet supportable, reliable and more importantly, be able to create a domestic ecosystem to support (not via Government-linked companies).

I did propose many times to create skunkworks projects and/or run hackathons to create solutions using open source tools to seed the next generation of local solutions providers. As I write this, it has not happened.

To compound the lack of thought leadership, the push in the 2000s to “outsource IT” meant that what remaining technically skilled people there were, got shortchanged as the work went to these contract providers (some of these skilled people were transferred to these outsourcee firms but left shortly after, because it was just BS).

This also meant that over time, the various entities who outsourced IT were just relationship managers with the outsourcee companies.It is not in the interest of the outsourcee companies to propose solutions that could lower the cost overall as it could affect the outsourcee’s revenue model. So, you have a catch-22: no in-house IT/architecture skills and no interest at all on the part of the outsourcees to propose a lower cost and perhaps better solutions.

I would be happy, if asked, to put together a set of solutions that will steadily address all of the healthcare IT requirements/solutions. I want this to then trigger the creation of a local ecosystem of companies that can drive these solutions not only for Singapore’s own consumption as well as to export it globally.

We have the smarts to do this. The technical community of open source developers are, I am very confident, able to rise to the challenge. We need political thought leadership to make this so.

Give me the new hospital in Woodlands to make the solutions work. I want to be able to do as much of it using commercially supported open source products (see this for a discussion of open source projects vs open source products), and build a whole suite of supportable open source solutions that are open to the whole world to innovate upon. It would be wonderful to see (no it does not exist yet).

There are plenty of ancient, leaky, and crufty systems in the current healthcare IT systems locally. We need to make a clean break from it for the future.

The Smart Nation begs for it. 

Dr Vivian Balakrishnan said the following at GeekCamp SG 2015 (video):

I believe in a #SmartNation with an open source society and immense opportunities; where everyone can share, participate and co-create solutions to enrich and improve quality of life for ourselves and our fellow Singaporeans.

And for completeness, the actual post is here (it is a public page; i.e., no account needed):

I am ready.  Please join me.

What I’ve learned in 15 years at Red Hat

[image credit:

In September 2003, I was asked by Gus Robertson to consider joining Red Hat – essentially to kickstart the Red Hat office in Singapore/ASEAN. I was at that time running my own open source consultancy – Maringo Tree Technologies.

I did try to get Red Hat to acquire MTT, but that did not go far. Red Hat had just done a JV in India and was not looking to do another even as an acquisition so soon. In any case that JV was subsequently bought out fully by Red Hat in 2005.

I started at Red Hat as the Chief Technology Architect. I would not say that Red Hat was my dream company to be at, but most everything I did before was all about Free and Open Source Software.

Time Before Red Hat (yes, there was)

I had no direct interest in the notions of software licensing (see this post about the Singapore Ohio Scientific Owers’ Group from 1982 – yes, predates establishment of the venerable Free Software Foundation). Even my 1989 MSEE thesis (link to PDF) was about creating software but I never did consider the licensing of it because lots of it was given to me or copied from USENET newsgroups.

I, together with Greg Hosler, Mathias Koeber, Ng Hak Beng,  and a bunch of others, set up the Singapore Linux Users’ Group in 1993. That was about when Red Hat itself was setup. I was fortunate enough to have been in situations where I could help drive the conversation and technology forward. I helped by distributing diskettes of Soft Landing Systems, Yggdrasil, Slackware etc either via the LUGS or directly. I was more concerned with getting the code out there and people benefitting from it. LUGS even created our own distribution in 2000 called EasyLinux.

I wrote up some how-tos – IP Masquerade, IP Alias, DialD, and lots of other stuff – I can’t recall what else – but suffice to say, these efforts were unexpectedly “rewarded” by shares being offered when Red Hat was planning on an Initial Public Offering.

I recall receiving an email from Red Hat and eTrade in 1999 (unfortunately, I can’t locate that email) in which it said that I was being given some shares (I don’t recall how many) to acknowledge the contributions made to the community. The only reference to the thinking at Red Hat about this is in an article in the Linux Magazine of November 15, 1999.

July 1999: The Community

It was clear that Red Hat wanted all the open source developers who had made its success possible to participate in its public offering. Red Hat would be nowhere without the hackers, and the company knew it.

Red Hat Director of Technical Projects Donnie Barnes spent three weeks scouring the Internet, digging up all the contributor lists to all the open source projects he could find. Red Hat then had to craft a letter to this list of developers. The SEC has a complex set of rules about what companies can and cannot say when they offer shares to the public. If a company doesn’t stay well within the rules, the SEC can –
and regularly does — withhold permission to proceed with an IPO.

“I’m sure they have very important and well-researched reasons for implementing each and every one of these rules,” Young said. “But to the companies who have to negotiate these rules on their way to a public offering, they appear designed solely for the purpose of ensuring the mental collapse of anyone who attempts to navigate through them.”

For example, the SEC-imposed quiet period “was one of the more bizarre notions to a salesman like me,” Young said. “How can you sell shares in your company if you are not allowed to promote your company for three months before your IPO or for a further month after your IPO? This letter to developers could describe the offer, but not mention any reason why anyone should want to accept the offer. That would be
promoting the shares in the quiet period — a big no-no according to the rules.” Red Hat ended up with a letter which, while legally acceptable, was “sufficiently badly worded to end up alienating a significant percentage of the developers we mailed it to,” Young said.

The SEC has a set of rules governing who is eligible to purchase shares in an IPO. First, you must be a U.S.-based taxpayer to buy IPO shares that are listed on an American exchange. This eliminated about half of the developers on the list from participating in the offer, according to Red Hat.

The SEC also has a set of rules designed to protect the public from scam artists who use public stock offers to con inexperienced investors out of their money.

“In effect the SEC deems IPO offers to be extremely high-risk investments, and therefore buyers of shares in IPOs must prove that they are experienced investors who can afford to lose the money they are being asked to invest,” Young said.

Unfortunately a significant percentage — about 15 percent — of the developers to whom Red Hat offered shares were either students or otherwise inexperienced investors by the SEC’s standards.

“And of course this offer was not being made by the SEC — it was being made by Red Hat and E*TRADE. So when members of the development community that we had extended the offer to found themselves declared ineligible, they initially naturally blamed Red Hat and E*TRADE,” Young said.

The final result was that well over one-fifth of the developers on the list were interested, eligible, and able to participate in the Red Hat IPO.

(extracted from

I accepted the shares from Red Hat and I guess I was part of the 1/5th.

For the record, I think I did sell the shares later (along with the shares that I later received from VA Linux when they IPOed). It was a very nice gesture on both Red Hat’s and VA Linux’s part and I am thankful.

When Red Hat offered me a position in 2003, I was very aware that this is a business that is trying to do the Right Thing.  While I was trying to figure out the value of such a move, I was provided with the relevant employment documents to sign. In one of them, a particular paragraph stood out for me.

In the Red Hat Code of Business Conduct and Ethics (a PDF from, this paragraph was pivotal:

Participation in an open source community project, whether maintained by the Company or by another commercial or non-commercial entity or organization, does not constitute a conflict of interest even where you may make a determination in the interest of the project that is adverse to the Company’s interests.
(from the bottom of page 2 of the PDF)

I did not, then, have to think twice. I signed.

With that stoke of the pen, my tenure at Red Hat started officially on 8th September 2003. I still had Maringo Tree Technologies and I had to find a way to exit it and eventually I did.

Fifteen years is a significant amount of one’s lifespan to be spent with a single entity (assuming a 30-40 year work career). Red Hat is the 7th organization I am with: CSA Group (1st job, employee, Singapore), Microsoft (intern, employee, USA), Sembawang Media (management Singapore), Brokat (technical management Singapore/Germany), Inquisitive Mind (co-founder/management/technical Singapore) and Maringo Tree Tech (founder/management Singapore),

Each stint provided experience and exposure to different aspects of the business/technology world. Perhaps the most challenging was to be your own boss at Maringo Tree Tech. While the time spent at MTT was relatively short, I am proud of what I was able to achieve. I suppose the work at MTT  was the lead-in into Red Hat for I had approached Red Hat as my own business even though, technically, I was an employee.

The thought process that says that the place you are employed is your own business, instead “it is just a job”, is a very powerful and empowering mental state. I want to make open source wildly successful and I was using Maringo Tree Tech as the vehicle to make that happen. When Red Hat came knocking on the door, it made immense sense to me that I should hitch on the Red Hat branding to drive my conviction forward.

I, had, at the back of my mind (and perhaps still valid today), that should I ever decide to part with Red Hat, I will still be doing the exact same thing but under a different label. That would be just fine.

What does 15 years at Red Hat mean for me? First, it is testimony that it is a growing organization and will continue to define how technology should be created, curated and consumed. That, the pursuit of fiscal goals has to be in congruence and cadence with ensuring the health and wealth of the commons. The commons comprise people from around the world who, once empowered by technology, are able to display and benefit from their talents and skills that can only but move the needle forward positively.

My early years at Red Hat were spent making sure the ONE product we had, Red Hat Enterprise Linux (2.1 and 3) were being made available to customers. I was tech support, sales, order entry, delivery person, spokesperson, evangelist, keynote speaker, community pointman and the go-to-for-everything-open-source-and-Red-Hat in South East Asia person.

It was not until late 2004 that Red Hat decided to re-draw the sales organization for Asia Pacific in creating an APAC HQ in Singapore for sales and marketing, Red Hat Brisbane to be the APAC engineering HQ, and to create six regions within APAC (ASEAN, ANZ, India, Korea, Japan, China). This change meant that we will have new Red Hatters focused on the selling of products and services but who might not have cut their teeth in free/open source software but were from the dark side, aka, proprietary software vendors. This is a necessary “evil”, but something that can be managed with the right amount of training and exposure. That was what happened when Red Hat University was setup in 2006 and I was asked to head that up for APAC, or more specifically the Sales College (which was the only college under RHU in APAC then).

I think, my time with the Sales College and especially in running the week-long Sales Boot Camp, was a very important personal challenge. Through the Sales Boot Camp, I helped transform a whole lot of open source migrants to become champions of open source. Not all of them got to their open source a-ha moment, but those who did, were sold on it. They could not do anything else as the fact that we are not shortchanging our customers by promises and handwaving but with solid and independently verifiable technology was not something to be walking away from.

In the meantime, our product portfolio started to grow. JBoss came on board in 2006, Qumranet in 2008, Makara and Gluster in 2010, CEPH in 2014, Ansible in 2015 and CoreOS in 2018 just to name some major acquisitions.

My sales quota in 2003 was US$100k and Red Hat’s revenue then was about US$126.1 million. Today, 2018, revenue is at US$2.9 billion. That’s a whopping 23x growth in revenue in 15 years. In the 60 quarters I’ve now been with Red Hat, every quarter has been better than the previous.

While the business grows with a plethora of products and service, where in the scheme of things does Red Hat understand how the open source community of projects are faring? These projects become open source products that our customers consume (via a subscription). The incredibly fine balance of open source projects and open source products is something that I’ve speak about because there is still an insufficient appreciation of what an open source project is when it become a product that a corporate entity has to be held accountable for.

Wtih the fine balance of project vs product, it is just as important to be able to gain insights into the open source projects that feed into the products. This was what lead me to start an internal project “Prospector” project in 2012. In 2017, Prospector became part of the CHAOSS Project of the Linux Foundation. Much work continues to be needed and I am working on many fronts to make that happen.

Which in a meandering way, brings me to today, 8th September 2018.

What would the next 15 years bring? What I know for sure is that it will be based on a widely distributed and decentralised world built on free/open source software and open hardware.

This will be a world that will bring the remaining unconnected 3+ billion members of the human family into the digital world, The tech flavours of today (AI, Machine Learning, Distributed Ledgers), having been built on free/open source software, will blaze the trail into new applications to better the human condition and the planet’s health.

The future is so open and bright, I need shades!

You gotta love the UNIX way of doing things

A colleague was asking for some help in a demo he was planning on that would involve him editing a spreadsheet in LibreOffice on his laptop and then the spreadsheet to be saved to a remote machine. Once the edit to the spreadsheet was done, that act should then trigger a process in the remote machine. He could do that by manually firing off a process, but would like it to be automated.

Because of his question, I found out that in LibreOffice, there is an option to “Save Remote” which gives me a range of options to work with:




I could then just save to a SSH-enabled destination. Pretty cool.

Turned out, for some reason, the LibreOffice version on his machine just did not successfully connect to the ssh destination. It worked out of the box on my Fedora 28 and RHEL 7.5 systems. Hopefully, a bug report will be filed.

So, as an alternative, I suggested that perhaps he should do an sshfs (dnf install fuse-sshfs -y) link to the remote system. This allows for the remote location to be accessible from all apps on the local machine.

The next thing to do was to write a script on the remote end that watches the spreadsheet for changes and to trigger some actions.

This is where the UNIX philosophy of simple, well defined and well written tools come into play. It is also a realization that there are a million ways to make it happen.

In this instance, I picked on entr (dnf install entr -y on Fedora and via EPEL or RPMFusion for RHEL) which does exactly what was needed and all in just one line. I did initially consider inotify but figured entr was simpler for the task at hand.

Here’s a screencast to show two terminals (in Gnome, ctrl-alt-shift-R):

r7: is a RHEL 7.5 system (on the left) and gillman: is a Fedora 28 system (on the right).

The line “ls file.csv | entr -p ./” is all that is bring done on r7. On gillman, I sshfs mounted r7:/home/harish onto gillman:/home/harish/mnt. When the file file.csv is edited on gillman, the file is saved via the sshfs mount point onto r7 and on r7, the one line command line executes via entr to trigger the shell script “” which is nothing more than:

r7$ cat

ping -c 5

The screencast is self explanatory.


My updated GPG keys: 0x61AFA27B

I have just updated my GPG keys.

The fingerprint is: 0x61AFA27B

And my public key is:



You can check my keys at any of the key servers. Here are two:



I am updating my keys to 4096 bits and that is the reason for doing it now.

You can also use if you prefer – this still has my previous key, and it will be updated to the new one.