Evaluate stack options for the v2

[antoinewolff]

The website (v1) use currently Wordpress + Divi.

This stack is problematic for differents reasons :

• Difficulty in updating the content (it is often an integrator who has to take care of it).
• Poor performance (loading time).
• SEO issues related to the sources page (#124)
• Rather long integration and development time for new parts.

The needs are:

• Let the integrators take care of the content from the start. These do not need a CMS.
• Maximize performance and eliminate all unnecessary frameworks.
• Find a solution allowing indexing of the sources page (#124)
• Quickly integrate version 2 by going directly through the code.
• Keep the current content of the pages.

The first ideas are:

• Directly code the front of version 2 in pure HTML/CSS/JS.
• Use a Headless CMS.

[antoinewolff]

@lutangar : Do you think we speak about this subjet this week ?

[gregoirelacoste]

Keep me in touch, i am very motivated about setting up a headless CMS. And maybe with strapi ? https://strapi.io/ And why not Gatsby ? https://www.gatsbyjs.com/

[antoinewolff]

Call 7/04/21 : @gregoirelacoste, @julienadamcom, @antoinewolff said :

Needs :

• On home : list a text inputs to update specific text parts (no structural changes)
• On inside pages : structured text that can be edited with a Markdown WYSIWYG (with stylized preview like on GitHub or/and with a "preview" button to render the page with AJAX and display it in a modal).

Stack considered :

• Strapi to manage the content : @gregoirelacoste has a good experience on it, @julienadamcom thinks it's a good idea but he wants to test it before.
• Gatsby to genrerate a static HTML/CSS/JS website : it use React, witch is a good thing if we want to work with common Storybook (extension + website) (it's on discussion with @newick).
• CleverCloud or Netlify to host the website.

Next step :

@julienadamcom will test Strapi and check Markdown WYSIWYG libraries. @gregoirelacoste will check what Strapi plugins can help us. @antoinewolff will schedule a new call for 9/04 at 4:30pm.

[lutangar]

Does it mean crafting the V2 from scratch is now officially on the roadmap?

Just wanted to mention my previous warnings here, apart from the substantial gain on the Developer Experience (and I know it matters) from the initially described issues:

Difficulty in updating the content (it is often an integrator who has to take care of it).

I thought @KaliviArf was now autonomous to edit the content.

2. Poor performance (loading time).

I didn't know it was an issue. What am I missing?

3. SEO issues related to the sources page (#124)

This won't solve this issue.

4. Rather long integration and development time for new parts.

So true.

So from where I stand, if we have a sufficient amount of work to do on the website to reach a critical milestone (to be defined) and the time spent on the new version will actually speed up the delivery of the critical features (to be defined) for this milestone then it need to be done asap otherwise it might not be a top priority.

[christpet]

@lutangar really good points. We are still in evaluation mode - we need to answer questions like these before we kick off officially.

I would love to set us up to address the SEO issue - @antoinewolff I know Pierrick has done some good thinking on this

[antoinewolff]

Call 12/04/21 : @gregoirelacoste, @julienadamcom, @JalilArfaoui, @lutangar, @antoinewolff said :

Important points :

• Markdown is the good way to format the page content.
• We must be sonverein on our data.

Solutions considered :

• Solution 1 : full HTML/CSS/JS
  • Faster on short term but not fitting with the dev team middle term strategy : mono-repo, shared components, storybook...
• Solution 2 : React + Edition on GitHub (or ~Gatsby only)
  • Beginner on GitHub is not optimal. GitHub is designed for developpers, not for content managers !
• Solution 3 : Strapi + Gatsby
  • A content managers friendly interface.
  • One additional server for Strapi with DB addon
  • No history ? (except if we find a way for DB on text file)
  • More maintenance

/!\ Might find a way to have best of both worlds and have md files in git as a source of truth

Conclusion :

Solution 3 seems to be the good one but we must check if :

• How much time ? :
• Who ? : Gregoire or Julien can be available.
• Concurence ? Do we have enough resource on the team... ? How do we organize the differents works ?
  • @antoinewolff/Marco/@newick : must finish the design.
  • ... ?
• Steps ?
  • Step 1 : develop the website with React
  • Step 2 : define on the mockups the data that can be updated in Strapi
  • Step 3 : design the data model in Strapi
  • Step 4 : plug the website on the strapi APIs

Next step :

• Call with @MaartenLMEM (with everybody if they want)
• Planification

[JalilArfaoui]

Some interesting Strapi plugins ?

• https://www.npmjs.com/package/strapi-plugin-versioning
• https://www.npmjs.com/package/strapi-provider-upload-github