The current top answer on stackoverflow is (in my not-so-humble opinion) incorrect. You do not need to import pandas or any other nonsense to convert a numpy datetime64 to a python datetime or vice versa.
Hey, wait a minute! this works fine. the pydatetime variable is a datetime.datetime object! That means the conversion to microseconds in the first example must be unnecessary!
…well, sometimes it is. But, lets try the same example, but this time with a more precise datetime object:
# Define our datetime64 object, this time with a nanosecond-level precision
import numpy as np
numpy_datetime = np.datetime64('2024-12-03T00:00:00.123456789')
# Convert to datetime.datetime
pydatetime = numpy_datetime.astype('O')
print(pydatetime.strftime('%Y-%m-%d'))
>>> Traceback (most recent call last):
>>> File "<stdin>", line 1, in <module>
>>> AttributeError: 'int' object has no attribute 'strftime'
Well, that didn’t go so great. Why is the pydatetime variable an integer now? The difference in behavior lies in the difference between datetime.datetime objects and numpy datetime64 objects and the reason they exist in the first place: numpy objects can have more precision, down to 1 nanosecond (1e-9 seconds). This makes them useful for certain scientific applications that require this level of precision. However, being a numpy dtype, they don’t have all the nice comforts of the builtin datetime objects like strftime and strptime. Since datetime objects can’t have nanosecond-level precision like datetime64 objects can, if you ask numpy to convert the datetime64 object to a “generic object type variable” (the ‘O’ dtype), numpy gives you an integer number of nanoseconds since 1 Jan 1970. This means that if you want to convert to a datetime object, you MUST give up precision beyond one microsecond (1e-6 second), and you must do this by explicitly casting to a microsecond-precision datetime64, then casting to datetime.datetime, as shown in the first example.
I have some concerns with how a WordPress, a *technically* open source project, depends on plugins (also open source, most of the time), that are hosted in a centralized repository. This repository can be taken over at anytime by Automattic. This itself is not really a concern for me, I was perfectly happy with my packages for the iOS jailbreak community (back when that was a thing—throwback!) being hosted by BigBoss and Chariz. Those were also controlled by dictators, but they were benevolent.
it’s clear to me that Automattic is not benevolent and will destroy the trust in open-source if you’re sponsored by the wrong people.
I’ll be switching this site over to “something different” when I get a chance. It will probably be Quarto as I’ve been using that for the documentation for xlma-python, but probably won’t be until the end of the semester.
On June 22, 2024, the Aggie Doppler Radar (ADRAD) was removed from the roof of the O&M building and was replaced by a currently-yet-to-be-named ClimaVision radar system. This was met with simultaneous fanfare and criticism. I ran ADRAD for 2 years and became more than a little emotionally invested in the system, and I’m an S-band and open source fanatic, so I was originally disappointed to learn of the agreement. However, after learning the details of the new system, I quickly changed my mind. This got quite long, so I’ve broken it up into sections, generally following the path of an electron through the radar chain.
This thread is entirely my personal opinion and does not reflect those of TAMU, ClimaVision, TTU, etc. Information gathered from public sources and Chris Nowotarski’s and Chris Weiss’ courses on radar meteorology.
ADRAD’s reflector removed by helicopter, June 22 2024
Transmitter
Summary: ADRAD has the advantage of longer wavelength/less attenuation, and longer range with higher nyquist velocity. ClimaVision has the advantage of dual-polarization and higher resolution. Differential Phase and longer pulse lengths can somewhat mitigate the attenuation issue, and the range/nyquist disadvantage can be almost completely mitigated with staggered/dual PRF techniques.
A magnetron, for reference. The coaxial output is to the right, and the glass bulb on top is a vacuum tube for control.
Design
ADRAD contains a single coaxial magnetron capable of producing a peak envelope power of 1 MW with a wavelength of 10.4 cm (S band). The ClimaVision radar has two separate solid-state transmitters capable of producing 500 W PEP each with a wavelength of 3.09-3.25 cm (X band).
The first obvious advantage of having separate Tx/Rx chains is dual pol. To keep this a shorter read for folks familiar with radar, I’ll collapse sections with deeper dives into the radar science that you can open if you want.
Dual pol explanation
The summary is that the two separate transmitters send out two separate waves, one of them with the electric field pointed horizontally, and the other pointed vertically. This allows for measuring the power returned (and therefore reflectivity) of volumes in two offset directions. By knowing the direction of the transmitted wave and the amount of power received in each direction, it is possible to derive the differential reflectivity (difference in the reflectance in horizontal and vertical beams, allows for the determinations of the shapes of targets) and correlation coefficient (the statistical correlation between the horizontal and vertical pulses sent to the same sample volume, allows for determination of the similarity of the targets in the region).
Additionally, by having two fully separate transmitters and not just two receivers (cough, NEXRAD), linear depolarization ratio (provides more information about shape and canting angle than just differential reflectivity) can also be recovered. The upshot of all this is, real microphysics research can be performed by students and faculty using the new system. I have personally used ADRAD in research presentations, and the limited (reflectivity, velocity, and spectrum width) variables available were a major downside. If I had access to the dual pol moments, specifically differential reflectivity and linear depolarization ratio, I think I could’ve answered more science questions in my projects, and I’m excited to see what future generations of students come up with.
ADRAD’s non-coherent transmitter
While the old magnetron allows for a higher transmit power, magnetrons are “phase-incoherent”. Physics details for why are attached below, for interested readers.
A model of a magnetron warming up, depicting the random “mode competition” process before the three electron spokes form.
Why Magnetrons are phase incoherent
Magnetrons are specially designed copper cylinders capped by two magnets with their poles facing the same direction. A central cathode receives a large voltage. The electrons in the central cathode are attracted and begin to flow outwards towards the surrounding anode, but are bent into curved paths by the flanking magnets. Due to random charge levels present on the anode, some parts of the outer ring are more attractive than others. The electrons flowing outwards are more dense in the more attractive regions. This has an effect on the outer anode, which is shaped in a way that it forms a miniature inductance-capacitance circuit. As the electrons from the center reach the outer edge, negative charge shifts from one side of the capacitor to the other. The inductors across the two sides of the capacitors delay this, leading to more clumping of the electrons flowing outwards until distinct “spokes” have formed, and the charges in the anode are oscillating at the desired frequency. However, for a brief period at the startup of the magnetron, the spokes have not formed yet. The time for the spokes to form and the oscillations to stabilize is based on the random distribution of charges when the process starts. As can be seen in the video, this only lasts for a few tens of nanoseconds, but this results in the magnetron never producing a consistent phase waveform between pulses.
The disadvantages of non-coherent transmitters are discussed in the following sections…
ClimaVision’s Coherent Transmitter, Pulse Compression, and Range Resolution
Something BIG is coming soon and we can’t wait to tell you what it is!
The new ClimaVision radar has a transmitter based on a phase coherent oscillator with solid-state amplification. This means that the outgoing pulse has a consistent and controllable phase. The first impact of this is pulse compression. By using two (or more) separate oscillators pre-amplification that are out of phase with each other, a pulse does not have to take the form of a standard sine or cosine wave. Binary data can be transmitted in the pulse, and can be differentiated in the received signal. Range resolution can therefore be increased by a factor of the number of bits transmitted. Usually, range resolution depends only on pulse length, and increasing range resolution results in lowering the pulse duration and therefore total energy transmitted. ClimaVision’s radar has a gate spacing of 32.15 meters. By assuming a standard 4-subpulse Barker encoding, this gives a pulse duration of 8.5 microseconds compared to ADRAD’s 0.7 microseconds, but because pulse compression is possible with a coherent transmitter, ClimaVision’s radar has nearly 4 times the range resolution.
It’s important to note that while transmitters typically measure their output POWER, the more important metric is the transmitted ENERGY (power times pulse duration).
Power vs. Energy explanation
ADRAD’s power is 1,000,000W compared to 500W, suggesting a 2000x more “powerful” radar, but given ClimaVision’s ability to use longer pulse durations without harming range resolution, this is not as much of an issue. It is not trivial to calculate the exact difference because I don’t have measurements of the power envelopes of either ADRAD or ClimaVision’s radars, but even assuming a square wave where both transmitters are immediately at full power when the pulse begins and maintain that power level throughout the entire pulse, this gives 0.7 Joules for ADRAD and 0.00425 Joules for ClimaVision… still much less, but only a roughly 150x multiple, and this is the worst case scenario for ClimaVision, magnetrons take longer to reach full power than transistors. I am also quite speculative of whether ADRAD actually still produces 1MW…
This goes a long way to combating the downside of lower transmit power, and also means that the ClimaVision radar uses less energy (as in, lower electric bill) than ADRAD.
Unfolding Second-Trip Echos with FM Chirp
Another advantage of phase coherency is second-trip echo identification. Pulse compression can again be used, instead of just including a binary datastream, the frequency of the pulse can change throughout the duration of the pulse. This frequency can vary pulse-to-pulse and be identified by the receiver. This is referred to as linear or non-linear FM chirp. If a received pulse is detected to have the previous pulse’s variation in frequency, it is therefore known to be a second trip echo, and the receiver HARDWARE can discard it or unfold it to the correct range. No fancy software algorithms, this is all done before a single line of code runs. This doesn’t completely solve the ‘Doppler dilemma’, but it makes the trade-off much less impactful. ADRAD, with its phase-incoherent transmitter, suffered from intense second-trip echoes, and filtering using a more rudimentary normalized coherent power thresholding resulted in the removal of some interesting meteorological data.
Differential Phase
Phase coherency also applies in the dual-pol realm! Combined, these two features allow for a product called “Differential Phase”, allowing for discerning liquid rainfall amounts, and is not subject to attenuation. This can also be used for research like the other dual-pol products, and can also correct attenuated reflectivity values.
This is actually the less advanced staggered PRF, not dual PRF
Solid-State transmitters also have the ability to change their Pulse Repetition Frequency (PRF) rapidly. The PRF is what determines the maximum unambiguous range velocity. The range folding issue can be handled using pulse compression, so to mitigate velocity aliasing… we have staggered and dual PRF. Staggered PRF is a strategy where two consecutive scans are performed at the same elevation angle, while dual PRF alternates between two PRFs between small groups of pulses. This allows using a low PRF to obtain reflectivity, differential reflectivity, and correlation coefficient at long ranges, and high PRF to obtain velocity, spectrum width, and differential phase values near the radar. Using pulse compression to range unfold second-trip echoes, velocities can be collected from the high PRF scan outside the typical max unambiguous range. The shorter X-band wavelength does reduce the maximum unambiguous velocity, but by comparing the aliased velocities between the low and high PRFs, it is possible to dealias the velocities by subtracting integer multiples of the nyquist velocity at each PRF from the measured velocity of each PRF until the values agree. Again, this is all doable with no gimmicky “feature detection” software algorithms, and can be done entirely using hardware logic. ADRAD was never capable of any of these techniques.
Maximum Range
The advertised maximum range of the new radar is “60 miles” which gives an operating PRF near 1550 Hz. When compared to ADRAD’s 300 Hz PRF with a range of 310 miles… yeah, ADRAD had much longer range. But consider the situation: ADRAD was built before the WSR-88D (and even the WSR-74C) networks. A longer range radar made sense at the time because otherwise, there was no coverage at all. But in 2024, things are different, NEXRAD and the TDWR networks do exist, and between KFWS in Dallas, KGRK in Granger, TIAH in Houston, and the Corsicana CORAD S-band, College Station has plenty of radar coverage, just not low-level coverage. The surrounding radars have plenty of coverage of the low-levels in their respective areas, so there’s no reason for a radar in College Station to see far by using a longer wavelength that the reflector can’t handle, which is a perfect transition into the antenna comparison.
Summary: ADRAD’s reflector was never appropriate for its wavelength, and ClimaVision’s narrow beamwidth will produce beautiful data. I’m not yet sure how to feel about the radome.
Design
Both ADRAD and ClimaVision’s antennas use a circular parabolic, axial feed antenna. Apart from the separate H and V channel feed horns on the ClimaVision antenna, there’s very little difference between the two designs.
Yesterday our team was in College Station, Texas for our media day with our newest university partners, Texas A&M! We had the honor of sharing more about our partnership and mission to cover low-level weather observation gaps with press covering the event.https://t.co/etbIerchQzpic.twitter.com/LGTmlHmqon
The diameter of ADRAD’s reflector was 4.88 meters. The transmit wavelength of 10.4 cm gave a 1.54 degree beamwidth. This is not good, even for a long-range surveillance radar. Overall antenna gain for ADRAD was 41.33dB. The ClimaVision system has a diameter of 2.44m, but with a transmit wavelength of 3.09cm, giving a beamwidth of 0.92 degrees, much more in line with NEXRAD. This provides much finer resolution detail of features. ADRAD’s longer wavelength was great for seeing far when that was necessary, but now that NEXRAD exists and it’s not a requirement to see so far, the large 1.54 degree beamwidth really was arguably ADRAD’s biggest downfall. The gain of ClimaVision’s antenna is more than a doubling on the linear scale, at 45.8 dB, again, compensating for the limited transmit power by increasing the Effective Isotropic Radiated Power.
Radome
Shockingly, the radome is a choice I’m not yet sure I agree with. A radome allows for the radar to run continuously, even during high wind situations like tropical cyclone landfalls (not unheard of in College Station, ADRAD had to be parked vertically several times, most recently during Hurricane Ike 2008, and at some point, the wind sheared the elevation brake off, damaging a large section of the rear waveguide). The radome will also lessen the need for maintenance of the pedestal/bearings, which is what killed ADRAD in the end. However, when the radome gets wet, this will increase the attenuation of the beam, and that stinks.
Scan Rate
ClimaVision’s smaller and protected antenna allows for a faster scan rate of up to 48 degrees per second, compared to ADRAD’s 18 degrees per second. A full PPI sweep now takes 7.5 seconds instead of 20 seconds. Especially given the focus the on low-levels, with little need to use higher elevations, this is an impressively fast update time. Even phased-array systems, which can steer the beam electroncially, are usually limited to a 60-90 degree azimuth range and must mechanically spin to capture a full PPI. While these systems are able to capture a full volume during a single rotation, a College Station based radar has no interest in the mid-/upper-levels of the atmosphere (they are already covered by NEXRAD), so I’m not even sure a phased-array system would justify its cost here.
Receiver
ADRAD’s receiver has a minimum detectable signal of -110 dBm and a dynamic range of 90 dB. ClimaVision’s has an MDS of -114 dBm (again, logarithmic scale, this means that the receiver can detect signals with less than half the power of ADRAD’s) and a dynamic range of 95 dB. Not much to say here, the new receiver is better in literally every way, and again contributes to making up for the lack of transmit power. Oh, and there are two of them, one for each channel.
Officially, I don’t know this for certain, but having talked to people who do, I will make an “educated guess” that the new ClimaVision receiver can detect a target with a reflectivity of 0 dBZ at a range of 40km. ADRAD may have once been capable of this, but especially with the state of its duplexer, it was never going to again.
In summary, I think that even if ADRAD worked as it did in the late 90s, the ClimaVision system makes more sense as a gap-filling radar between the nearby NEXRAD sites.
Data Availability
Alright, fine, I’ll get to the controversial bit. There are essentially four classes of data users for a radar in College Station: – TAMU faculty and students – Local media – National Weather Service – Former students, weather enthusiasts, etc.
Both ADRAD and ClimaVision provided Level 2 data in real time to TAMU faculty and students. With the new ClimaVision system, TAMU users will now be able to conduct and publish more research with it (dual pol!), so I think this is a major advantage for them. I’m not sure of the exact limitations on how long the embargo on publishing data or research using the data is, what (if any) the limitations are for how much data can be published, etc. TAMU students will have access to the entire ClimaVision network, not just their own, and I can see the one in Hamilton being useful for TASC chases in Gatesville.
ADRAD was never able to provide data to local media. Technically if a station had the media version of GR2Analyst, it could, but there was never a way to guarantee its reliability to a degree that a station would’ve liked, and ADRAD was only run very infrequently. While I was at Texas A&M, I did try to work the technical side of creating a WSI Max compatible feed, but only made very limited progress, and the politics of publishing this feed was never really worked out. I don’t think it ever would’ve been possible. With the ClimaVision system, the radar will be run 24/7, and data will flow in real time to any media station that buys it, with support, and all of this is professionally managed infrastructure, not managed by an undergrad. I personally have no issue with ClimaVision wanting to be compensated (radars are expensive… as TAMU learned over the last 51 years…), especially from media who are using the data commercially.
ADRAD was instrumental to the National Weather Service over the years. The most obvious case of this was March 21/22 of 2022, when I pulled an all-nighter to run the system during a tornado outbreak in the area. The power lines that supplied the KGRK 88D had been run over by a tornado earlier in the day, and ADRAD managed to run for over 24 hours straight. The Weather Service later thanked the department for running the radar, and had issued several warnings based on our data. The new ClimaVision system is reportedly accessible to the NWS via Synoptic Data/the National Mesonet Program. I have heard conflicting reports as to how easy it is for weather service forecasters to obtain the data. For what it’s worth, I am currently trying to open an edu-associated account with Synoptic Data, but the free tier has significantly less features than the commercial/NOAA tiers, and I’m not sure if I will get the radar data or not.
Unlike ADRAD, ClimaVision does not let just anyone have access to the raw, ‘Level 2’ data feed for GR2Analyst or py-ART or whatever your favorite way of viewing those was. I can completely understand why this sucks, as I fall into this group personally (pending the edu account with Synoptic, which I don’t expect to work out). The only thing I can really say here is “ADRAD was never going to publish data again regardless”. Many efforts were made by the department writing proposals to many organizations to try to get ADRAD or a replacement funded, but sadly no one wanted to fund it. Considering that… the weather enthusiasts lost the L2 feed for the College Station radar in spring of 2023 when ADRAD died. The ClimaVision system didn’t kill the GR2 feed, it just exposed the feed’s death. If you fall into this category, I’d recommend following KBTX weather. The entire team are A&M former students, and they have access to the ClimaVision data. It’s not the same as getting to play with it in GR2 in real time, but you can still follow local area weather there, and I think the added uptime of the radar is worth this trade off.
So, the trade off is, we weather enthusiasts lost access to the data, but the local media stations get immediate access, 24/7. As much as it personally sucks… I think the ClimaVision system will have more reach and potential for the protection of life and property than ADRAD did.
Does managing conda environments hang over your head like a dark cloud? Is more of your research time spent waiting for Anaconda Navigator to open than doing the actual research? Don’t worry–there is an escape from this torture! mamba!
For the uninitiated, this all has to do with python package management. Python is a fantastic programming language, but primarily because of the many many great packages you can install and use through python. There are packages for EVERYTHING, I use python for radar and lightningmeteorology, but researchers of all fields from signal processing to astronomy to molecular biologists all have niche packages designed by their communities, for their communities, to do their research with.
To install these packages, you need a package manager. Python comes with its own built in package manager called ‘pip’, which is fine enough, but pip installs all packages globally. This is cool, but sometimes you want to switch back and forth between multiple versions of packages or different sets of packages. Pip is not capable of this.
Anaconda, and Miniconda, are package managers developed by Anaconda, Inc. that have a concept of different ‘environments’ that the user can switch between quickly. Anaconda comes with a GUI and a set of pre-installed pacakges that are guaranteed to work with each other, making it user friendly. Miniconda does not, making it more flexible.
But these have some limitations. The first is the above screenshot. Anaconda in particular is bad about taking FOREVER to search through your installed environment (it has to process all the packages you currently have installed to determine what version of the package you want to fetch, and since Anaconda comes with a set of preinstalled packages, it has to look through all of those), and a large/old enough miniconda environment can do the same. Additionally, Anaconda Inc’s licensing does not allow free, commercialized use of the suite.
mamba is a C++ drop in replacement for conda, except faster. You can install it into an existing anaconda or miniconda base environment, or, if the licensing is an issue or if your base environment is completely broken, completely remove (ana/mini)conda and install mambaforge.
If you want to install it into an existing base environment (this is not recommended by the developers of mamba) conda install -n base --override-channels -c conda-forge mamba 'python_abi=*=*cp*'
If you want to remove your current conda installation and replace it with mambaforge: Windows: 1. Open windows settings->apps->scroll to miniconda or anaconda and remove it 1a. If you selected “add miniconda python to PATH” during installation, you will need to remove this manually. Search “env” in the start menu, select “Edit the system environment variables”->”Environment variables…” and delete all keys relating to conda. 2. Download the mambaforge windows installer 3. Proceed with installation as described in the install instructions for windows 4. When running conda commands, replace “conda” with “mamba” and enjoy the extra time!
macOS/Linux: 1. Open a Terminal of your choice 2. Type in conda config --show-sources 3. Delete all paths listed in the output of the above command 4. Type conda info and make a note of the “base environment” path. In the example below, it’s /Users/vm/miniconda3
5. Type conda init --reverse --all 6. Delete the base environment path from step 4 7. Exit your terminal and restart 8. Type curl -LO "https://github.com/conda-forge/miniforge/releases/latest/download/Mambaforge-$(uname)-$(uname -m).sh" && bash Mambaforge-$(uname)-$(uname -m).sh 9. When running conda commands, replace “conda” with “mamba” and enjoy the extra time!
So in the most “meta” post possible, the most interesting thing I’ve done recently is create this website. I wanted to do it the fun way, entirely self-hosted. I feel like the days of hosting your own website are numbered, with DDoS attacks on the rise and the ever-cheapening of hosting services like AWS, etc. (please don’t attack this site — I definitely don’t have the bandwidth to handle it). It honestly probably would’ve been cheaper to just use a hosting service, but I also wanted the “fun” of building a computer.
This site is run on a system powered by the Ryzen 7950X… which yes, all I’m doing right now is running a website on it. I very easily could do this on a raspberry pi (if those weren’t unobtainium), but some day, after I graduate, I’d like to do post processing of weather models and observations. My language of choice for this is python, which is… inefficient to say the least. So I’ve taken the “just throw more hardware at it” approach. I got a free RAM kit from Micro Center which was nice, and the CPU was already discounted for holiday sale or something. I only took two pictures of the build itself, oh well.
as I’m writing this I realize I have installed the heat sinks with the Noctua logos upside down 🙁
Matplotlib is unfortunately entirely CPU-bound. As such I do not have a GPU in this system.
The next fun challenge was picking a domain. wx4stg.com was an obvious choice. – The number in the name and the seemingly random combination of letters that don’t spell anything (just like t-a-m-c doesn’t spell anything either…) makes it cheaper. – However, it’s my ham radio callsign, so that “random” combination of letters and number doesn’t have any significance… except to me. It’s unlikely that anyone will want to buy it off me later.
I chose Google domains because I want G-suite email and all the other services. Yes I know this is not privacy friendly. I would much rather use iCloud mail but Apple have decided that “if you trash an email, it gets automatically deleted after 30 days”, which I didn’t know about, for years… so I was trashing emails thinking that I could go back to them as long as I never emptied the trash… nope. And in Apple’s signature “we know better”, there’s no way to turn off that limitation. I would really much rather just host an email server myself, but after searching around, it seems that all of the big email providers “spam filter” out anything that comes from a self-hosted email server… so it seems I’m too late to the party there.
As for software, the system is running Ubuntu Linux, specifically the Server 22.04 LTS channel. I picked this 1) I first “learned” on a jailbroken iPhone, and jailbreak bootstraps have always been some derivative of the debian software stack. I know how the apt package manager works and have run an apt repository before. yum/rpm/dnf/pacman/etc. are much more unknown to me. 2) it’s popular and everyone else does it.
If you somehow came to this page because you want to set up your own website, I essentially just followed this tutorial for LAMP and another for phpmyadmin by digitalocean. I have some experience (Texas A&M HDWX) with managing a LAMP stack, but I had never built one from zero before, and those guides got me up and running within ~15 minutes. (if anyone from digitalocean is reading this–you guys rock for making those tutorials and updating them)
Then the fun part, actually getting traffic to my computer. When I was ~14, I had self-hosted a Minecraft server for my friends (this was before hosting providers for Minecraft were really a thing, and 14-year old me would’ve had no money or card or paypal to pay for it anyway, so self-hosted was the only option. I ran it on a very old, even for the time, laptop that I also played Minecraft on at the same time… needless to say it wasn’t exactly the best experience, but…) it got me experience with IPs and ports from a very young age. Unfortunately I live in an apartment complex in College Station. Solution: just run the server at my parents house. Port forwarding was almost straightforward, my parents have Comcast Xfinity internet. The Xfinity app is the only way to port forward these days, you can no longer just allow a port from the local http server hosted by the router, which has positives and negatives:
+ I can open a port on their router from ANYWHERE, even without being on network. This gives some additional security as I can, from college station, decide that I need SSH access, open port 22, log in, make whatever changes I need, and then close the port for security. – If the app has a bug, which it did, that prevents me from adding a port forward, then the only option is to factory reset the router.
Anyway, opening ports 80 (http) and 443 (https) was easy enough. Yes I did say https, and yes if you look in your browser’s address bar right now, there should be an HTTPS lock present. I had always heard that this was a pain and cost money to buy an SSL cert from a trusted CA, but huge thanks to Let’s Encrypt and the Electronic Frontier Foundation for making the process take < 5 mins.
This leads to the not-so-fun part, binding “wx4stg.com” to my parents’ public IP address. Shout-out to Justin at Xfinity support for walking me through how to do this, but essentially there are two routes: buying an IPv4 static IP ($$$) or using dynamic DNS (free). I chose the latter, of course. As I understand it, Comcast (like everyone) is conserving their IPv4 space, and somehow it’s more efficient for them to change your public address, randomly? I honestly don’t see how this saves any space. But since a computer can google (or ask an API…) “whats my public IP” and always be aware of its public IP (even if it’s dynamic), that same computer can then inform google domains of changes to the public address, and google can update their DNS records. I’m not sure if this will be sustainable forever, but for now that’s what I’m using. One caveat if you’re reading this like a tutorial, ddclient’s google domains support didn’t work for me as it was trying to update an A record with an IPv6 address, which you … can’t do. I had to configure ddclient for an “other” service:
Finally, I’m using wordpress.org (the cool HTML/PHP software for self-hosting a blog) which is different than wordpress.com (which is itself a hosting service). I’m not yet entirely sure if I like wordpress or not. I really really wish there were something like Apple’s AutoLayout that could create HTML/php. But oh well. For now I’m not yet hosting weather products here (I’ll continue to leach TAMU’s server power until they tell me no), but when that happens, that’ll almost certainly require custom HTML/php.
Thats all for now I guess. If anyone is actually following this like a tutorial and I forgot something, send me an email.
