\
**Category:** Ecosystem\
**Created:** [January 9, 2021, 11:00am UTC](https://discuss.ocaml.org/t/generating-html-tyxml-vs-mustache/7065 "2021-01-09T11:00:57Z")\
**Posts on this page:** 8\
**Page:** 1
**Author:**  [@lindig](https://discuss.ocaml.org/u/lindig)\
**Post date:** [January 9, 2021, 11:00am UTC](https://discuss.ocaml.org/t/generating-html-tyxml-vs-mustache/7065/1 "2021-01-09T11:00:57Z")
I’ve developed a tool that generates a report in HTML by using [OCaml Mustache](https://github.com/rgrinberg/ocaml-mustache) and found that working well. Now I’m contemplating turning this into a web service using [Opium](https://github.com/rgrinberg/opium). I notice the preference of Opium for TyXML and understand the safety and control it provides. However, it is quite a step away from a template based approach and maintaining a lot of HTML inside OCaml source code does not appeal to me. Any experiences?
---
**Author:**  [@Khady](https://discuss.ocaml.org/u/Khady)\
**Post date:** [January 9, 2021, 11:31am UTC](https://discuss.ocaml.org/t/generating-html-tyxml-vs-mustache/7065/2 "2021-01-09T11:31:43Z")
There’s [TyXML JSX syntax](https://ocsigen.org/tyxml/4.4.0/manual/jsx). It gives a template like system with a very close to html syntax + some type safety. I find it convenient to use. The possible downside of this solution is that it’s using the reason syntax which might not be here forever. I don’t have a significant experience with mustache to compare.
---
**Author:**  [@Armael](https://discuss.ocaml.org/u/Armael)\
**Post date:** [January 9, 2021, 11:58am UTC](https://discuss.ocaml.org/t/generating-html-tyxml-vs-mustache/7065/3 "2021-01-09T11:58:09Z")
This will not be immediately useful to you, but it’s worth noting that @Drup had at some point the project of integrating a template engine with tyxml, exactly for the purpose you mention ([Choice of template language for tyxml](https://discuss.ocaml.org/t/choice-of-template-language-for-tyxml/1863)). And some very recent work by @gasche on ocaml-mustache (in particular [https://github.com/rgrinberg/ocaml-mustache/pull/58](https://github.com/rgrinberg/ocaml-mustache/pull/58)) looks like it could be very useful for that project.
---
**Author:**  [@shonfeder](https://discuss.ocaml.org/u/shonfeder)\
**Post date:** [January 9, 2021, 12:39pm UTC](https://discuss.ocaml.org/t/generating-html-tyxml-vs-mustache/7065/4 "2021-01-09T12:39:24Z")
There’s no need to fuss with Reason, right? You can just use [TyXML Ppx syntax extension](https://ocsigen.org/tyxml/4.4.0/manual/ppx)
---
**Author:**  [@mro](https://discuss.ocaml.org/u/mro)\
**Post date:** [January 9, 2021, 3:44pm UTC](https://discuss.ocaml.org/t/generating-html-tyxml-vs-mustache/7065/5 "2021-01-09T15:44:32Z")
At times I prefer to “purely” separate raw data and presentation. So I publish domain specific xml plus an xslt to html
e.g. [https://demo.mro.name/geohash.cgi/u28br](https://demo.mro.name/geohash.cgi/u28br) or
[https://demo.mro.name/shaarligo](https://demo.mro.name/shaarligo) or
[http://rec.mro.name/stations/b2/now](http://rec.mro.name/stations/b2/now)
So there’s very little markup-fiddling in code (which I don’t find worth while a dependency) and all markup stuff is in the browser in xslt - no javascript involved.
Main Downside: Google won’t index neither the resulting html nor the raw data.
---
**Author:**  [@lindig](https://discuss.ocaml.org/u/lindig)\
**Post date:** [January 9, 2021, 5:01pm UTC](https://discuss.ocaml.org/t/generating-html-tyxml-vs-mustache/7065/6 "2021-01-09T17:01:05Z")
I think the main advantage of a template-based approach is that it works well in these situations:
- design prototyping and turning a design into an application
- using an existing design framework
- collaboration between designer and developer
Keeping the entire design in TyXML creates a barrier. I also like that Mustache has a CLI that populates a template from JSON data, which supports prototyping.
---
**Author:**  [@shonfeder](https://discuss.ocaml.org/u/shonfeder)\
**Post date:** [January 10, 2021, 1:33am UTC](https://discuss.ocaml.org/t/generating-html-tyxml-vs-mustache/7065/7 "2021-01-10T01:33:18Z")
I just came across this old thread, which you might find interesting: [Choice of template language for tyxml](https://discuss.ocaml.org/t/choice-of-template-language-for-tyxml/1863)
---
**Author:**  [@Drup](https://discuss.ocaml.org/u/Drup)\
**Post date:** [January 10, 2021, 6:50pm UTC](https://discuss.ocaml.org/t/generating-html-tyxml-vs-mustache/7065/8 "2021-01-10T18:50:51Z")
> [@lindig](#):
>
> I think the main advantage of a template-based approach is that it works well in these situations:
>
> - design prototyping and turning a design into an application
> - using an existing design framework
> - collaboration between designer and developer
I entirely agree, and I think it’s very important to enable the “traditional” workflow of teams building web applications. That’s why I wanted to have a tool that would take a standard “templated HTML” file, and turn it into an OCaml module using tyxml to build the HTML.
Unfortunately, time is limited and it’s a complicated piece of code to write, so I stopped working on it. I think now all the pieces are there, so it’s time to finally do it. As @armael said, mustache’s recent additions make it a valid templating language for this job, just need to write the tyxml side. 🙂
I think I’ll free some times in the coming month to work on that, but any help is appreciated, either in tyxml directly, or to improve the mustache OCaml implementation! 👍