|
I'm a mobile developer and at work there are several executives who are very interested in developing PWAs (Progressive Web Apps) for enterprise apps instead of Native. By that I mean for use in mobile browsers, not a hybrid framework like Flutter. My general impression is that the experience associated with PWAs is that they're convenient but lovely. If I only wanted to use the app rarely I might use the PWA version but if I planned to use it regularly, for example every day at work, I'd want to use a native version. However I'm a bit biased and I've heard that younger users don't mind using PWAs in a mobile browser even if they plan to use the app regularly. What are people's thoughts on this and is the thing about younger users preferring PWAs accurate?
|
# ? Dec 15, 2023 00:05 |
|
|
# ? May 31, 2024 02:03 |
|
SaTaMaS posted:I'm a mobile developer and at work there are several executives who are very interested in developing PWAs (Progressive Web Apps) for enterprise apps instead of Native. By that I mean for use in mobile browsers, not a hybrid framework like Flutter. My general impression is that the experience associated with PWAs is that they're convenient but lovely. If I only wanted to use the app rarely I might use the PWA version but if I planned to use it regularly, for example every day at work, I'd want to use a native version. However I'm a bit biased and I've heard that younger users don't mind using PWAs in a mobile browser even if they plan to use the app regularly. What are people's thoughts on this and is the thing about younger users preferring PWAs accurate? Are you including things like wrappers using tools like Ionic in PWAs?
|
# ? Dec 15, 2023 00:43 |
|
mystes posted:I wouldn't think true PWAs are being used widely enough yet for it to even be possible for younger users to prefer PWAs (I would imagine that most people haven't used a single PWA still?) Ionic can build both PWAs and hybrid apps, but I'm specifically talking about PWAs used via a browser. Apparently dating apps and ride hailing apps are pretty popular PWA use cases.
|
# ? Dec 15, 2023 01:04 |
|
SaTaMaS posted:Ionic can build both PWAs and hybrid apps, but I'm specifically talking about PWAs used via a browser. Apparently dating apps and ride hailing apps are pretty popular PWA use cases. Which ride hailing apps? The major ones are definitely not PWAs.
|
# ? Dec 15, 2023 01:21 |
|
OK, so it looks like uber's mobile site m.uber.com is a pwa in that it has a pwa manifest and a service worker which does the bare minimum of caching, but not even push notifications or anything like that. I'm guessing that's what's being referred to. However, I don't think that anyone, at least in the US, is using that rather than Uber's native apps. I think to the extent that most sites are using PWAs at all right now, it's like that: if you really want to, you can install their mobile site as a PWA just to get the icon on your phone, but it doesn't really offer any actual PWA functionality. That said, if by "enterprise app" you mean something for internal use in a company, offering what's essentially just a way of getting an icon for a website, with no particular additional functionality, might be sufficient? If you want to make something that can actually work offline or offer other additional functionality over a normal mobile website, I think there are still enough problems with PWAs that it's probably not worth the trouble compared to using some other method. I'm also curious about where the information about the popularity of PWAs is coming from. Googling you can find various sites making claims about how common PWAs are, but I would be interest in actual numbers about how many people are using PWAs. mystes fucked around with this message at 01:55 on Dec 15, 2023 |
# ? Dec 15, 2023 01:47 |
|
mystes posted:OK, so it looks like uber's mobile site m.uber.com is a pwa in that it has a pwa manifest and a service worker which does the bare minimum of caching, but not even push notifications or anything like that. I'm guessing that's what's being referred to. I use the PWA for Outlook because using the full app requires me to enroll in MDM and I don't want to. It still supports push notifications and everything else I need, so it's perfectly fine for me.
|
# ? Dec 15, 2023 01:56 |
|
Does the Google productivity suite (Gmail, drive, sheets/docs/etc) count as PWAs? I haven't really tested how functional they are offline, but I don't think it's zero.
|
# ? Dec 15, 2023 02:00 |
|
Computer viking posted:Does the Google productivity suite (Gmail, drive, sheets/docs/etc) count as PWAs? I haven't really tested how functional they are offline, but I don't think it's zero. They seem to have some sort of PWA that you can install but when I tried it just now it didn't seem to work offline at all. That might have just been a glitch, but the existing offline functionality that worked in chrome used a chrome extension or chrome app (the latter being functionality that google deprecated a while ago telling everyone to use PWAs instead, but since PWAs still aren't great in various ways, it wouldn't surprise me if google is still using chrome apps for its own software). mystes fucked around with this message at 02:12 on Dec 15, 2023 |
# ? Dec 15, 2023 02:03 |
|
Is there a thread related to xml stuff?
|
# ? Dec 21, 2023 04:46 |
|
goodness posted:Is there a thread related to xml stuff?
|
# ? Dec 21, 2023 04:55 |
|
I realise that with XPath and XSLT there are aspects of XML that are like whole programming languages in their own right, but no it doesn't have its own thread. It probably wouldn't get enough activity to stay visible. Just post in here
|
# ? Dec 21, 2023 20:06 |
|
Hammerite posted:I realise that with XPath and XSLT there are aspects of XML that are like whole programming languages in their own right, but no it doesn't have its own thread. It probably wouldn't get enough activity to stay visible. Just post in here From wikipedia, XSLT is Turing complete.
|
# ? Dec 24, 2023 19:20 |
|
kaschei posted:From wikipedia, XSLT is Turing complete. Bored grad students are the best and worst parts of society
|
# ? Dec 24, 2023 19:29 |
|
Volmarias posted:Bored grad students are the best and worst parts of society The worst was when they did this with SPARQL and now those grad students are data scientists who insist on everything being a graph with no design constraints.
|
# ? Dec 24, 2023 19:39 |
|
kaschei posted:From wikipedia, XSLT is Turing complete. having used it, the only suprising thing here is that it took til xslt version 2 to become so
|
# ? Dec 25, 2023 00:45 |
|
Hammerite posted:I realise that with XPath and XSLT there are aspects of XML that are like whole programming languages in their own right, but no it doesn't have its own thread. It probably wouldn't get enough activity to stay visible. Just post in here xml features are mostly stupid bullshit that you shouldn't use anyway - just parsing an xml document is complicated and potentially a security vulnerability and has all sorts of bullshit going on with different character sets, etc, and you have to cringe whenever a developer has the audacity to try to create xml through string concentration because that's a recipe for summoning eldrich horrors. dtd are useless because in the beginning, the people who developed xml thought that it was possible for actual users to edit xml files without loving poo poo up (i have spent literal years in zoom calls guiding people through editing xml files, no they cannot), and that it would be nice to have arbitrary regex run on the created document to make sure these manually created xml files would have all the right info - but in the real world, if a document fails your dtd/schema definition, no one is going to have a clue what your definitions mean anyway or how it relates to your application, so that's loving useless. as for other xml technologies, one thing i have also noticed is that software developers as a rule are lazy assholes that will do anything to avoid modeling their data in a relational database, and poo poo like xslt and xpath queries are technologies that enable that kind of bullshit - like i've seen people who do this for a living put giant strings of xml in database columns and it makes me very sad. as for xml signatures and xml encryption, there's no reason that poo poo needs to be built into the standard imo, that's solving the problem at the wrong layer.
|
# ? Dec 25, 2023 02:06 |
|
Bruegels Fuckbooks posted:one thing i have also noticed is that software developers as a rule are lazy assholes that will do anything to avoid Yep with you so far, this is correct quote:modeling their data in a relational database, and poo poo like xslt and xpath queries are technologies that enable that kind of bullshit - like i've seen people who do this for a living put giant strings of xml in database columns and it makes me very sad. The inner platform pattern is An Anti Pattern, but I would be remiss if I couldn't find excuses for this. Changing the database may be a dramatically more costly action while still experimenting, the data doesn't actually need to be in a relational database because it doesn't need to be (or can't realistically be) compared, gently caress poo poo dick making GBS threads nipples damnit I've got 30 minutes to write this thing or I'm fired because Very Expensive Client will drop us, this is the exact content we were given by endpoint A and we don't need to transform it before sending it to endpoint B, etc. The developer is probably, but not necessarily, a stupid idiot hell fucker here. quote:as for xml signatures and xml encryption, there's no reason that poo poo needs to be built into the standard imo, that's solving the problem at the wrong layer. I assume most of the XML extension stuff, like XSLT, is from the 2000s, when people still thought it was going somewhere. It's ugly as sin, but as an exchangeable data type it beat the hell out of everything else at the time and had the complexity to be able to handle most models reasonably well. Inevitably, time marches forward, and the problems are laid bare for all to see. But, without the benefit of hindsight letting us know the flaws, and with an understanding of how companies want to "extend" standards with useless garbage they can sell people, it's pretty easy to see why it was so eagerly embraced.
|
# ? Dec 25, 2023 04:45 |
|
Volmarias posted:I assume most of the XML extension stuff, like XSLT, is from the 2000s, when people still thought it was going somewhere. It's ugly as sin, but as an exchangeable data type it beat the hell out of everything else at the time and had the complexity to be able to handle most models reasonably well. Unfortunately I have been doing this since XML was new, but I dealt with it just because SOAP was the way to do web services back in the day and having a standard for serializing/deserialization seemed better than the alternative. I was always extremely apprehensive about the complexity of xml and didn't realize just how much bullshit I was in for when the xml fad hit the company I was working for, and it literally has wasted years of my life. JSON does everything that I wanted xml to do with much less bullshit.
|
# ? Dec 25, 2023 10:14 |
|
Bruegels Fuckbooks posted:xml features are mostly stupid bullshit that you shouldn't use anyway - just parsing an xml document is complicated and potentially a security vulnerability and has all sorts of bullshit going on with different character sets, etc, and you have to cringe whenever a developer has the audacity to try to create xml through string concentration because that's a recipe for summoning eldrich horrors. dtd are useless because in the beginning, the people who developed xml thought that it was possible for actual users to edit xml files without loving poo poo up (i have spent literal years in zoom calls guiding people through editing xml files, no they cannot), and that it would be nice to have arbitrary regex run on the created document to make sure these manually created xml files would have all the right info - but in the real world, if a document fails your dtd/schema definition, no one is going to have a clue what your definitions mean anyway or how it relates to your application, so that's loving useless. as for other xml technologies, one thing i have also noticed is that software developers as a rule are lazy assholes that will do anything to avoid modeling their data in a relational database, and poo poo like xslt and xpath queries are technologies that enable that kind of bullshit - like i've seen people who do this for a living put giant strings of xml in database columns and it makes me very sad. as for xml signatures and xml encryption, there's no reason that poo poo needs to be built into the standard imo, that's solving the problem at the wrong layer. [sitting in an armchair, writing on a notepad, talking to Buegels Fuckbooks who is reclined on a couch] And how did that make you feel?
|
# ? Dec 26, 2023 12:23 |
|
Bruegels Fuckbooks posted:xml features are mostly stupid bullshit that you shouldn't use anyway - just parsing an xml document is complicated and potentially a security vulnerability and has all sorts of bullshit going on with different character sets, etc, and you have to cringe whenever a developer has the audacity to try to create xml through string concentration because that's a recipe for summoning eldrich horrors. dtd are useless because in the beginning, the people who developed xml thought that it was possible for actual users to edit xml files without loving poo poo up (i have spent literal years in zoom calls guiding people through editing xml files, no they cannot), and that it would be nice to have arbitrary regex run on the created document to make sure these manually created xml files would have all the right info - but in the real world, if a document fails your dtd/schema definition, no one is going to have a clue what your definitions mean anyway or how it relates to your application, so that's loving useless. as for other xml technologies, one thing i have also noticed is that software developers as a rule are lazy assholes that will do anything to avoid modeling their data in a relational database, and poo poo like xslt and xpath queries are technologies that enable that kind of bullshit - like i've seen people who do this for a living put giant strings of xml in database columns and it makes me very sad. as for xml signatures and xml encryption, there's no reason that poo poo needs to be built into the standard imo, that's solving the problem at the wrong layer. Did you time travel here from the year 2004
|
# ? Dec 26, 2023 16:27 |
|
XML. When you absolutely, positively got to trigger the fight, flight or gently caress response in every motherfucker in the repo. Accept no substitutes.
|
# ? Dec 26, 2023 16:34 |
|
So I had never really used xml before starting my current job where I'm part of a large migration to Workday. The part I'm doing is creating all the data reports and integrations, and quite often I'll use xsl to transform a workday report to fit the outbound destinations format requirements. Its a lot simpler than wrangling it in the more 'advanced' Studio editor they have in Workday. I had really only used WD as an employee before this, but its actually pretty easy to use on the administration side (I know it gets a lot of hate from users because of bad configurations). Previously I was an application developer at IBM mostly doing javascript and python stuff, then got interested in moving to some sort of Data position. Still looking for that but these WD careers pay surprisingly well and scale up pretty high. Especially you get into consulting or working for WD itself.
|
# ? Dec 26, 2023 19:39 |
|
in a just world we'd be living through the dream of web 2.0, creating cross-site mashups via well-typed WSDL schemas and XSL transforms
|
# ? Dec 26, 2023 19:53 |
|
I feel like if there hadn't been such a strong backlash against xslt we would have a lot more convenient tools to replace stuff like xslt by now and using them wouldn't seem so annoying
|
# ? Dec 26, 2023 20:18 |
|
God help me, but I kind of like XSLT. It's kind of relevant that I've only used it sporadically, no idea how much I'd hate it if it was my full-time job.
|
# ? Dec 26, 2023 20:37 |
|
Computer viking posted:God help me, but I kind of like XSLT. You can write horrors in it, but that's true of all programming technologies. If you can guarantee the schema of your input it works fairly well. The failure as ever is assuming xml can/should be hand edited.
|
# ? Dec 26, 2023 20:59 |
its still better than JSON
|
|
# ? Dec 26, 2023 23:07 |
|
The first-level problem with XSLT is that it’s a full-blown programming language using XML as its basic syntax, which makes it awful to both write and read without a dedicated tool. That tends to mask any other flaws it has as a language, which I can’t speak to very well because I’ve barely used it. XML overall has a lot of trade-offs which are pretty well-trod. It has a ton of tooling, but actually using any of that tooling generally has surprisingly high costs. Like, if you want your editor to validate your XML, you need to give it a DTD reference, but that will often also force your primary client to open a million extra network connections.
|
# ? Dec 26, 2023 23:08 |
|
If I have to consume the output of another program, gently caress yeah, I prefer that to be an xml. Because I know it will have a contract written (be it DTD or schema). I will know everything about it from said contract. There will be no surprises. JSON, though, is preferable when I have to hand edit things, as it's less verbose. But God help you if you have no documentation of said document, what can it have, what does it accept, etc. There is json schema now, but I only saw it used in vcpkg.json . It's nice when you have it.
|
# ? Dec 27, 2023 00:49 |
|
xpath is real good, and every imitator (css, whatever flavour of the month for json) should have just copied it wholesale
|
# ? Dec 27, 2023 03:34 |
|
Volguus posted:If I have to consume the output of another program, gently caress yeah, I prefer that to be an xml. Because I know it will have a contract written (be it DTD or schema). I will know everything about it from said contract. There will be no surprises. Hope the XML it outputs isn't some janky hand rolled library stuff where you get to fix it before you can even use it
|
# ? Dec 27, 2023 03:39 |
|
Volmarias posted:Hope the XML it outputs isn't some janky hand rolled library stuff where you get to fix it before you can even use it Then that's not xml, is some other concoction. To be fair, absolute junk emitted by programs I've only seen when said output wanted to pretend it's json. Then again, I haven't used xml in probably a decade or more. I used it in the late 90s and early 2000s. I may just have rose-tinted glasses memory.
|
# ? Dec 27, 2023 04:05 |
|
Volguus posted:If I have to consume the output of another program, gently caress yeah, I prefer that to be an xml. Because I know it will have a contract written (be it DTD or schema). I will know everything about it from said contract. There will be no surprises.
|
# ? Dec 27, 2023 12:41 |
|
I genuinely love XSLT, but making it a purely functional programming language with no way to loop other than recursion was an idea that sounds great from a PL theory perspective but was never going to be popular. XSLT 2 made significant steps towards being a language designed to actually be used but came along too late.
|
# ? Dec 27, 2023 18:01 |
|
Ok ok, if we wanna be pedantic today: "The odds of an XML (produced by a released application not written by drunken monkeys) having a contract are almost 100%. The odds of a JSON having a contract, and not just whatever properties the developer thought about that morning, are almost 0%." There, better? Yes I know XMLs can be written and well-formed without a contract, but I don't remember seeing one except during development of said application. Anyway, speaking of JSON schemas, I found another JSON that has (theoretically) a schema: CMakePresets.json. From the documentation page we find the helpful link to its schema. Nice, I thought. Adding the $schema property to the JSON to help the IDE give me nice auto-complete: quote:cmake --list-presets You were so close CMake. So close. So far vcpkg.json remains the only one that actually works.
|
# ? Dec 30, 2023 00:16 |
|
Volguus posted:Ok ok, if we wanna be pedantic today: "The odds of an XML (produced by a released application not written by drunken monkeys) having a contract are almost 100%. The odds of a JSON having a contract, and not just whatever properties the developer thought about that morning, are almost 0%."
|
# ? Dec 30, 2023 01:23 |
|
My condolences.
|
# ? Dec 30, 2023 01:27 |
|
Volguus posted:Ok ok, if we wanna be pedantic today: "The odds of an XML (produced by a released application not written by drunken monkeys) having a contract are almost 100%. The odds of a JSON having a contract, and not just whatever properties the developer thought about that morning, are almost 0%." I literally work for a company that makes an XML database and the number of customers I've worked with that actually write schemas for their XML documents is a rounding error. I'm lucky if they have declared namespaces.
|
# ? Dec 30, 2023 01:42 |
|
KillHour posted:I literally work for a company that makes an XML database and the number of customers I've worked with that actually write schemas for their XML documents is a rounding error. I'm lucky if they have declared namespaces. Volguus posted:My condolences. Hey, at least you can safely say that their application was written by drunken monkeys. With JSON, odds are that they were sober. They knew what they were doing. And that's even more frightening.
|
# ? Dec 30, 2023 01:54 |
|
|
# ? May 31, 2024 02:03 |
|
Volguus posted:Hey, at least you can safely say that their application was written by drunken monkeys. With JSON, odds are that they were sober. They knew what they were doing. And that's even more frightening. no tech, be it XML or json, can prevent a sufficiently determined/inexperienced dev from making GBS threads all over their API and or building the stupidest poo poo imaginable and somehow the places filled with these devs have no controls in place to prevent it, funny how that works (that's all places btw)
|
# ? Dec 30, 2023 02:05 |