The author of this presentation informs in this thread that his statement about Web 2.0 benefits is an opinion piece and not based on tests and measurements.
https://forum.xojo.com/t/web-2-0-benefits-more-simultaneous-users/75561
There have been several comments on TOF today questioning some of the presentations, the one on this thread is very accurate and honest. Wonder how long it lasts before itâs either visibly removed (deleted by author) or removed without traceâŠ.
What I am questioning is Xojo Incâs criteria for presentations (havenât seen any) and their overall sense of quality.
I wonder if David should have written âits EXPECTEDâ to support more users than 1.0 because of the changes made
Thats the impression Iâm left with after reading @Stretch327 comments on that thread
But as David notes its not based on empirical evidence of two equivalent implementations of âthe same appâ in Web 1.0 and Web 2.0 being benchmarked against each other
The impressive Part is: not based on experience. That would be so that I would speak about Vaadin, CN1, Jpro and so on without ever testing it and without using it by self. That is painful and should not be done while nobody can believe you after.
Comments are starting to turn nasty with âad hominem attackâ bing tossed in
Deletions should ensue any time now
I do agree with Ivan on his comments on HTTP/1.1. HTTP/1.x is known to have a slow response time. It does not efficiently utilize traffic by compressing the request and response headers and is ineffective in resource prioritization. HTTP/2 was well used and tested by the time they released Web 2.0. Their reasoning to not yet implement it because it was not âstableâ enough were ridiculous, especially with Xojoâs reputation of releasing unstable software all the time. Now HTTP/3 is starting to roll out on most browsers already, so calling Web 2.0 a âmodernâ framework is bit far fetched and will only work as âcommercialâ blabbering for folks who do not really follow the evolutions of the Web.
Same goes for using an already outdated Bootstrap framework at release. Bootstrap 5 was definitely already stable enough and well tested by then. The risk the problem would be in the (bad) implementation of it by Xojo was far greater than the problem being in Bootstrap 5 itself.
Being able to handle 50 concurrent users is nothing. Even a embedded jetty (java) server can easily handle thousands of them, as can PHP and most other web servers I suspect.
On first public release of Web 2.0 I did a couple of blog posts (here and here) on it and warned them they had designed a very bad event system that would cause a lot of trouble but they did nothing to optimize it, nor looked into what was sent to the browser (lots of CSS/JS was sent that was completely unused by the WebApp). The whole thing had a very amateurish feel when one looked under the hood and it was clear the âEngineersâ had little to no experience with writing WebApps.
I somewhat can agree win Greg that concurrent users depend of the WebApp. HOWEVER, there are plenty of ways to do some generic stress testing on your framework that CAN give you some indicational numbers you can provide to your users:
e.g.
- Simple handler that writes some text as the response
- Handler that connects to a local MySQL database and issues all kind of queries
- Handlers that write and manipulate Web pages small/medium/heavy loaded with HTML components
Then you can use one of the many stress testing tools out there and find out where the breaking point is. A good one will cost plenty of money (which Xojo Inc probably doesnât has) but as a Web Server vendor, at least you would KNOW how good or bad your product is and you can stop posting those wild guesses and stupid analogies (the cow thing was hilarious!).
I bet they did do such tests internally (I would be surprised if they didnât), but that the numbers werenât that great and not for the public eye. Especially if they wanted to be able to sell at least a couple of Web 2.0 subscriptionsâŠ
AFAICT, the official product documentation does not provide any information about performance or scalability, nor does it have links to documents that could provide such information.
Buy it and find found at your own expensesâŠ
There is a product documentation but no product specification.
In the case of cars, manufacturers have always tended to set specifications slightly higher or better. Xojo, however, doesnât seem to succeed in this, either due to a lack of interest or some other reason.
Regardless, their reference app is quite underwhelming. By that, I donât mean that not even a preview can be generated, as it would be technically possible with every tool that I am aware of today.
This could slow be because Iâm accessing it from Europe, and itâs likely running on a cheap isolated server in the USA. In any case, the app should struggle if multiple users access it simultaneously.
As far as I can remember, GP once mentioned on TOF that at least 200 simultaneous user connections should be possible. However, they couldnât substantiate this claim at the time, and the community pushed back for obvious reasons. Regrettably, I can no longer find this post on TOF.
In any case, the browserâs debugger and the embedded network and performance analysis are already sufficient to identify the overhead of data transfers. Itâs an improvement over Web1, but certainly not groundbreaking in 2023 when compared to truly modern frameworks and technologies.
Not having implemented HTTP/2 at GoLive was a joke back then, but now itâs dangerously playing with fire.
Slightly off-topic, but as itâs a classic, Iâm more than happy to share this post once more.
https://forum.xojo.com/t/upcoming-webinar-developing-web-apps-with-xojo/75203
Genius doesnât talk about migrating Web1 to Web2 apps.
With all these webinars heâs pursuing his career as a genially gifted teacher.
Geoff was a trainer for 4D a long long time ago - but dont flatter him with _genially gifted _
And deep down I think he still has a desire for Xojo to BE a lot more like 4D (something I dont think I would shoot for as it was a quirky tool set)
If Xojo were more âFilemakerâ - drag a db field onto a layout and bingo bingo its connected and you can write some verification code & all the DB interaction is handled ⊠but with a ârealâ general purpose programming language
maybe
Oh, thatâs his self-styling, as we know
I just added a dose of irony
silly me
You donât have access to Bavarian beer anymore?
I am back in Canada so have to make do with what I can get here
I can send you LöwenbrÀu 5l can
Funny enough one German client tried to send me a bottle of alcohol and the courier refused to send it as they couldnât be sure the recipient was > 18 years old when they accepted it for delivery
So you might not be able to
Sadness ensues
Have to admit that one of my favourites is the old MĂ€rzen Ocktoberfest beer
The Festbier is ok but I like the other better
Oktoberfest MĂ€rzen is specially brewed for this event
Well that style - I know other breweries make beers that are similar and just drank some in Weilheim at a very nice little brewery
And at Ischgl
Yes it is Bavarian MĂ€rzen. Always nice