Oxygene: A modern language built on the foundation of Object Pascal | Hacker News Reader

Oxygene: A modern language built on the foundation of Object Pascal

remobjects.com

99 pointsluismedel33 comments

Ouch, those prices though. It's 100% their right to choose how they want to distribute their language, of course! But there's no plausible scenario where I'd ever write with a proprietary language implementation, let alone one I had to pay for. There are too many free/Free options to voluntarily lock myself in for any of the kinds of things I ever work on.

It's the Delphi pitfall :(

Still the best IDE I've ever used, but that was because in 1990s Ukraine everything was pirated. Wouldn't be able to use it any other way.

($1k per named developer the first year, $750/yr for renewals.)

They may know that the realistic target market is a few hundred licences, tops.

Lazarus is a pretty good open source alternative, except the documentation, and the tools to build said documentation, suck.

As for the price, inflation adjusting $50 in 1983 to today is $160, so the personal use price isn't horrible.

I don't have a big problem with the money: it's a fraction of month's salary. The usual problem with commercial languages is the freedom to set things up how I want and spin up new environments quickly. Mathematica, for instance, is a significant pain to run on my laptop, desktop, and a cloud machine.

I think there's room to innovate in licensing. I would err on the side of making it easy to install at the cost of some piracy. A great language could 100x its market share by losing 1/2 to piracy, so 50x overall.

It is competing with Delphi, which has about the same price

https://www.embarcadero.com/app-development-tools-store/delp...

I am not opposed to paid languages, but my expectations for a paid product is much higher than a free one. A lot of things I am perfectly fine with from a free language would be unacceptable from a paid one.

What I'm most curious about is how they get new developers in the first place? Nobody starting their career is going to think "I need to become proficient in a language, how about the obscure one I have to pay $1000 for".

They can't have much of a marketing spend or I would expect to have heard of them. Who is their audience? As an old Turbo Pascal programmer, I'm not asking that rhetorically, I'm genuinely curious.

I feel like the better "model" would have been to make a free and open source language with some structure to lead it (even could have been in house only like sql or something) and then create highly specialized tooling to go with the language sort of like Microsoft or Jetbrains does. I don't really condone this either, but at least it would have a chance to get off the ground.

Its a reallyyyy hard sell to pay for a language just so that you can create software, especially something that sounds scary like Pascal. (idk anything about Pascal specifically but thats the vibe I get and I guess people will feel similarly) what happens if I build something with a proprietary language and the rug gets pulled out from underneath me? Thats the kind of concern people should have. Thats a LOT of trust before even starting programming. Especially when so many great languages with great ecosystems already exist for free and extensive examples and docs.

For those wondering why are they charging per developer etc. This case study helped me understand what they actually are selling: https://www.remobjects.com/elements/casestudies/curacaoweath...

Their landing page and generally content on their page is terrible but maybe nobody stumbles on their site randomly looking for this. People are reacting to their pricing after seeing the title but I dont think that's what they are selling...

I'm not really sure how that link helps explain it since it's about other other languages that they sell, but the TLDR of this language is that it's cross platform Delphi. Although Delphi is cross platform Delphi, so I'm still confused.

I am at once compelled by the sweet tone of Pascal and recoil in site of mercantile connivance. In truth I do honor the allegiance to business practices of old. I did in my time employ compilers from cheerful Centerline and stout StepStone. Yet it is the winter of our commercial compiler discontent. Though I tremble at the visage of Wirth's ghost, redime te captum quam quas minimo.

(sorry, been reading Shakespeare recently.)

It's nice to see there's still a market for a commercial compilers, I'm not sure there's enough benefit for my work to justify the price tag. Would love to hear from people using it what they think of it. And it reminds me a bit of Modula-3, or at least where Modula-3 seemed to be headed.

They also have their own swift implementation apparently? That's the first time I see an alternative implementation of it. https://www.remobjects.com/elements/silver/

It's also for various platforms, so you can write Swift code comiling to a .Net executable say.

I think that this is really what Elements are selling the alternative versions of languages - not just Swift and Pascal but Java C# and VB.Net - and the ecosystem to go with them on all major non free platforms (ie not Linux or BSD)

Well, that was under the radar. I used Object Pascal back in the MacApp days and figured it had disappeared forever. And Microsoft is allegedly a customer? I’d like to hear more about that…

I never trust those "these are our clients" claims unless they tell me about a project. That could easily be "one guy with a microsoft.com email bought a license once".

It's just I've seen people doing it!

I wouldn't even frame that as "not trust": of course it's rarely more than that one guy! And technically, that's perfectly sufficient for what they claim, even though practically it's several orders of magnitude less than what a favorable reader might read into it. Take it with a spoonful or two of salt. This is not really different from "not trust", only a little less bitter.

Well, yes, it's technically correct, the best kind of correct ;) But it's tru it 's a bit mischaracterizing, making it look like they're kind of "official" in some departments there.

Sometimes they just link companies they "work with" meaning that often the relationship is the inverse of what they are trying to imply. Maybe they use Windows, so they claim to have a relationship with Microsoft..

Won't it make more sense to go with the Embarcadero implementation of Object Pascal (Delphi)? That is available free for users as long as you make less than USD 5000/year using that tool

It’s worth noting that this is commercial software that costs money - even for academic purposes - but I assume that the access to support and the full-time support engineers it funds is worth is for some business customers.

Too bad there's no details on the "modern" features vs Object Pascal. Or are they hidden?

Anyway, seems interesting. I gather that their business model is to find large institutional customers looking for a way to scale out their legacy code (Delphi, etc.) to new platforms?

I wonder where they find developers willing to learn their tools when the entry price to learn is non-free. Also, a closed ecosystem nowadays means depriving the devs of the rich ecosystems that have grown around open languages. Which can be a good thing... (looking at npm, ahem)

The new features are hidden in the documentation https://docs.elementscompiler.com/Oxygene/Delphi/NewFeatures...

Given that it doesn't support vcl I don't see a reason to ever move to it for existing delphi applications which I assume are mostly fat clients.

It's funny that they named "nil"lables "nullable".

I have often thought that there's scope for a descendant of Object Pascal to take on a new name to coalesce all of the new ideas while freeing up the namespace occupied by legacy features.

This is not it.

I especially like the ensure old syntax, https://docs.elementscompiler.com/Concepts/ClassContracts/

    ensure
      Count = old Count +1 
      : 'MyObject: Count logic error';

Never seen that before, but smartpascal also added it. Edit: blush, Bertrand Meyer introduced that already. Pretty

I find Bertrand Meyer's Effiel System very compelling. The design by contract methodology is baked into the language. I did try a DBC library [1] for Python but eventually gave up trying to build it into my work. Not because there is anything wrong with that library, but the extra syntax "mess" that results from the bolt on nature that the library provides to Python in order to work became too much for me in the end.

[1] https://github.com/Parquery/icontract

Has anyone tried how large are the binaries built with it, e.g. a Windows exe with a single window and label displaying just hello world? (release mode, 64bit)

Photoshop was originally written in Object Pascal, then transpiled to C and then C++.

If Apple brings back the Macintosh Toolbox, I’ll get right on porting it back.

I would never develop software on top of a proprietary language that I have no control over. The only way I would consider using a closed source language would be if there was a source available contract with some sort of "you own it perpetually if we ever stop supporting it" clause. (Who knows, maybe they have that)

not really a language but more of a Delphi competition. Delphi is still alive and kicking so there's a market for RAD for sure.

> Download Trial

Rest in peace.

Reply on news.ycombinator.com