search

MicrosoftDocs / azure-docs

Open source documentation of Microsoft Azure
https://docs.microsoft.com/azure
Creative Commons Attribution 4.0 International
10.31k stars 21.49k forks source link

OWL2DTDL Converter code sample provided in the article is no longer acitvely maintained #108398

Closed LeelaRajesh-Sayana closed 1 year ago

LeelaRajesh-Sayana commented 1 year ago

[Enter feedback here] The OWL2DTDL Converter code sample provided in the article is no longer actively maintained by the author. Furthermore, the converter does not check/validate the limitations of the DTDL V2 content limitations. This causes issues to the customers during model upload. It would hard to back track every issue in different models and fix them independently.

If the converter code throws a warning during the template conversion, this issue would be resolved.

Here is an issue reported from the customer https://learn.microsoft.com/en-us/answers/questions/1229967/error-in-loading-models-in-azure-digitial-twins

Document Details

⚠ Do not edit this section. It is required for learn.microsoft.com ➟ GitHub issue linking.

LeelaRajesh-Sayana commented 1 year ago

Hi @baanders , I appreciate it if you can take a look at the above feedback from the customer on the OWL2DTDL Converter

baanders commented 1 year ago

@hammar Can you take a look at this customer issue with the OWL2DTDL converter sample? Is this an update we can implement—or, otherwise, how would you recommend advising customers to avoid getting tripped up like this?

hammar commented 1 year ago

As I'm tasked with entirely different things as of late, I'd have to redirect questions for such tooling development and support to my management chain.

For context and as additional info: this tool was developed prior to my joining the company, for a limited input set (it was built to support the RealEstateCore project ontology and the subset of OWL it uses), and was never intended to be a product as such. At some point, we onboarded it to this repo because it seemed useful to others, but it has no product promise, spec, or warranty of any kind. Getting it to a state where it would support the entirety of the (quite complex) OWL spec, and guarantee correctness of DTDL output in all cases, is likely a very significant engineering effort.

Given this, unless someone wants to fund a major effort to evolve and maintain this tool, I'd propose managing customer expectations is a better mitigation strategy.

If someone does want to fund such an effort, and if my management chain agrees to it, I'd be happy to take part.

hammar commented 1 year ago

Re. managing customer expectations, I would propose framing this not as a production ready tool, but as a sample code base, from which those customers can take inspiration and code to develop their own ontology ingestion pipelines.

baanders commented 1 year ago

Thank you @hammar for the additional context and thoughtful response! That makes sense to me.

I'm adding some of this language to the doc where we introduce this sample, in this PR: https://github.com/MicrosoftDocs/azure-docs-pr/pull/235369.

After clarifying that the sample is not a comprehensive solution and suggesting that customers can take ideas and starting code from it to develop their own solutions, I think it's reasonable that it would not provide full error-checking. Thus, I'm going to have Git automatically close this issue when the PR with the additional language is merged.

However, @LeelaRajesh-Sayana feel free to continue the conversation here if needed.

Thank you!

LeelaRajesh-Sayana commented 1 year ago

Thank you @hammar for providing your insights on this. Rephrasing the wording around the usage of the OWL converter sets the right expectations to the customer. That would help address this issue.

Thank you @baanders for making the changes. I do not have any further questions on this.