About Michel Rodriguez
I am a freelance developer working on a range of projects.
My main area of expertize is XML, and technical publishing in general. I have served for a long time as the resident XML expert for a standard organisation.
Beyond this I also develop a range of project often dealing with data. The data is usually either difficult to obtain (Web Scraping), heterogenous (Document Management System) or incomplete (Machine Learning) and requires effort and creativity to be turned into a useful output.
I am the author of XML::Twig, one of the most popular modules for processing XML, and of a handful of other modules mostly XML/HTML related that can be found in my (meta)CPAN directory.
“Data is a precious thing and will last longer than the systems themselves.”
The types of services I offer :
- XML related: from document modelization to conversion to XML, XML QC, integrating XML data with other sources or clients, to general XML processing. Very good knowledge of XML for standards (STS, JATS) but also of MathML, more than 20 years of experience in the field of XML for techpub.
- Data cleanup and improvement: converting various formats to XML, PDF, JSON, CSV or loading them into data bases, adding structure to data (XML or other), dealing with character encodings, improving incomplete data using machine learning.
- Web micro-services: building REST APIs on top of various services, like machine learning or exposing XML repositories.
- SaaS: complete modern web systems, mobile-ready, using Bootstrap, hosted in the cloud if appropriate.
- Web Scraping: collecting and structuring publicly accessible data from websites, including dynamic websites.
Links
Recent Projects
- SaaS: management tool for the network of external experts for a consulting firm, including
full-text search of common types of documents, email alerts...
stack: linux, apache + starman, perl, dancer2, DBIx::Class, Text::XSlate, postgresql, bootstrap, jquery, postmark
challenges: deal with a range of document formats, manage automatic email interactions with users, manage all the aspects of SaaS - XML Quality Checking Tool: allows automatic checking of standards for a range of rules:
numbering, structure restrictions, style rules... the tool allows the creation of custom rules,
visualization of error context in the original XML or in HTML or PDF created from the XML, a wiki for
teams to share know-how about common errors... Its current targets are the STS and the S1000D but other
formats can be added.
stack: linux, apache + starman, perl, dancer2, SQLite, XML::Twig
challenges: allow creation of new rules in an easy way, report problems in a helpful way - Request Tracker (RT) participating in the development of RT 5. New features,
"coring" existing extensions, integration with external services, custom development for clients, support.
stack: perl, mason, DBIx::SearchBuilder
challenge: dealing with an existing complex software, deliver clean code, understand customers requirements and implement them within the RT framework. - Machine Learning: categorization of tender notices, assigning a CPV (Common Procurement Vocabulary)
code based on their title. Implemented as a micro service returning results in JSON
stack: linux, starman, perl, TensorFlow, grocery, Lingua::Stem, dancer2
challenge: deal with the quality of the training data, monitor and adjust the behaviour of the system - Web Scraping: scraping of a range of public websites for tender information, clean up the
data, normalize it and output it as XML to feed a search engine
stack: linux, perl, phantomJS/CasperJS, puppeteer, SQLite, XML::Twig
challenges: deal with a variety of website technologies, add structure to the data, monitor activity to detect changes in websites structure and aadjust processing -
mif2mml: tool converting FrameMaker's MIF equations into MathML (mif2mml on github).
stack: perl, Parse::RecDescent, XML::Twig
challenges: cover the entire MIF specification, based on Adobe's docs, deal with quirks in the way FrameMaker generates equations. -
Open Source: maintenance of the XML::Twig Perl module, widely used to process XML data in
Perl
challenges: make sure the module installs and passes the tests on a wide variety of configuration, help users and keep it relevant.
I am somewhat active on StackOverflow, under the moniker mirod, usually answering questions about XML::Twig. - Standards Dictionary: on-line tool for querying a dictionary built from the definition section
of all IEEE Standards.
stack: linux, apache + starman, perl, dancer2, SQLite
challenges: gather and QC the XML data, and interface the system with the standards management tools. - XML expert for the IEEE Standards Department: adaptation of the STS (Standards Tag Suite) DTD
to the IEEE standards
challenge: making sure all of the necessary data structure was captured by the STS/JATS based DTD used by the internal client, and that the XML could be used to generate standards with the same quality as before using an XML toolchain
Clients
More
Modern Perl is a set of modern conventions, tools and methods that effective Perl programmers use to write powerful, maintainable, scalable and concise code. It relies on CPAN, a distributed repository of thousands of Open-Source modules. Modern Perl favors OO, with a powerful ORM on top of the data base. The Perl culture is also very focused on testing and offers frameworks and numerous tools for building and maintaining tests.
I have been part of the Perl community for quite a while now, and I have followed the evolution of the language and its culture
I can bring you clear, documented, tested code, that uses as much as possible existing libraries to limit the amount of new code that needs to be written. This allows me to be extremely efficient and to deliver results in a timely fashion.
“Easy things should be easy, and hard things should be possible.”
I can help you with the whole life-cycle of your documents: help you design your DTDs, go through the data to check that it can be modeled properly, help with the conversion process, advise on choosing or designing an editing system and help you use the XML data to create new products.
“My definition of an expert in any field is a person who knows enough about what's really going on to be scared.”
It helps you manage your customer interactions by storing all the exchanges both with the customer and within the support team in one place. It is also allows extremely customizable, to fit your process and integrate with your other systems.
From installation to developing custom extension, I can help you with all aspects of RT deployment.
I have a very good knowledge of the software, what it can do out of the box and how to get it to do what it can not.
“If it’s not in the ticket, it didn’t happen.”
Machine learning is an exciting field that promises a lot. Can it deliver for you?
I have a good amount of experience with short-text classification, which often helps add missing information to data, based on the text that's available. It can be a simple and (relatively) easy way to improve the quality of the data, and deals fairly well with data sets that are noisy and of less than ideal quality.
“When you’re fundraising, it’s AI / When you’re hiring, it’s ML / When you’re implementing, it’s linear regression / When you’re debugging, it’s printf()”
“It’s automation, not automagic.”
“The expectations of life depend upon diligence; the mechanic that would perfect his work must first sharpen his tools.”
“I think, fundamentally, open source does tend to be more stable software. It's the right way to do things.”