Hello, I am MichaelJachan. I would like to work with u on HED800 / 3G.

[MichaelJachan]

Dear HED Community!

Ramon (I guess u know him) said, I can comment the code in here. Unfortunately (?), I have created a PullRequest (PR) with a very small code change (formulated as TODO-Comments)

If there is a way to annotate code (as if I were the PR-reviewer), I would like to use it!

BR and BTW, I like HED,

MJ:)

PS: I have reviewed the HED800.xml and i guess i have found 10-20 small issues in the codefile ... :)

[MichaelJachan]

I have some suggestions 4u, the authors (if i were a code reviewer) about the 800 XML file.

[MichaelJachan]

2\. We never directly edit the XML.  Our current process is that we edit the mediawiki in hed-specification/hedwiki and use tools to generate

Yes, these "tools to generate" is the place, where your programmer can care for a nice formatting, ie, indent in the xml.

[MichaelJachan]

8. IDEA: Would you like to consider providing separate numeric types for int and double (ie, float).

For now, I am not able to distinguish btw. int and double, when I read your xml.

Thank you :)

[MichaelJachan]

9. Why does the entry "/Property/Organizational-property/Event-context", which is the only entry with the "unique" attribute, does not have a # subtag?

[MichaelJachan]

4\. I have invited you to be a member. I will go through your suggests and add them to PROPOSED.m.  Future pull requests should make changes to PROPOSED.md.

Is it OK for you, if i keep on posting my comments here, instead of writing them into the PROPOSED file? I think, someone of your internal team shall take and preselect my issues from here and put them in there ...

HOWEVER, I would like to become a reviewer of one of your pull-requests :)

[MichaelJachan]

10. Warnings in HED-MATLAB-TOOLS:

WARNING: for plugin "HEDTools" version in the folder name "3.1.0" and in the eegplugin_ file "3.0.0" differ

EEGLAB: adding "HEDTools" v3.1.0 (see >> help eegplugin_hedtools) - new version 2.7.0 available

[VisLab]

This probably would be a good place to explain why the top-level schema changed between versions 8.0.0-beta.5 and 8.0.0.

Before the release, we did an intensive review with three aims in mind:

1. Making sure that the current structure of the HED schema did not preclude overlaying a HED ontology for future linked data work.
2. Making sure that every term stood on its own --- that is if you saw the word without its hierarchy you would not misinterpret what it meant.
3. There weren't any terms in the schema that were obvious mistakes or in the wrong place.

Results:

• The review took longer than anticipated and ended up delaying the release of HED-3G to the beginning of August.

• The need to be compatible with a future ontology (an BFO-based OBO Foundry ontology, to be specific) triggered a top level reorganization. The main difficulty was that OBO has some requirements about relations, which were not satisfied in 8.0.0-beta. We ended up changing the top-level (Agent-property, Data-Property, Task-property, and Attribute) to (Property, and Relation).
  Note: As long as annotators use the recommended short-form, moving terms around doesn't affect the annotation, so this isn't as draconian as it might seem.

• There were some problematic individual terms that changed. For example, Left-side has become Left-side-of to make it clearly a Relation.
  Another problematic tag was Visual which became Visual-presentation. In general Sensory has been separated into subcategories Sensory-attribute and Sensory-presentation under Sensory-property. This was probably the change that will affect previously annotated data the most, but it cleaned up a lot of clarity issues.

[VisLab]

The requireChild attribute is still under discussion. The issue is whether xxx should always require a value if xxx/# is in the schema.

An xxx tag that has a # child cannot be extended, even if its parent has an extensionAllowed. This restriction makes unambiguous downstream parsing possible.

However, the issue is whether xxx should be required to have a value. Our current thinking is to relax this requirement so that xxx can be grouped with tags from the Data-value sub-tree. This issue has not been settled, so the attribute has been left in 8.0.0, but for the most part is not used.

[MichaelJachan]

(I have deleted my old answer-post, because it was nonsense. Now I understand requireChild much better)

Before i can fully understand the syntax of requireChild, i need to ask another, more basic, question. I have asked Ramon the same, already:

Hed is a list of Tags, which can also be seen as a tree. Right?
A Tree has a root node, from which many branching nodes originate. Nodes are under other nodes. In the end of each branch, a node without any child nodes is called a leaf. Right?

==> So my Question is: Is any node from the tree a valid HedTag, OR are only the leave nodes valid HadTags?

[VisLab]

Every node is a valid HED tag. There is no requirement that it be a leaf node. You can use any node and tools will convert to its full path to the root of its subtree. The HED xml has as the root of the hierarchy. There is no HED in the tag path, that is just an XML thing. Thus, the HED hierarchy is actually a set of trees with roots: Event, Item, Action, Agent, Property, and Relation.

[VisLab]

A node with the tagGroup attribute means that it must appear in some tag group (has parentheses around it somewhere). We are probably going to put the Relation subtree in the category when we start doing more sophisticated processing of relations. A topLevelTagGroup means the node has to be in a tagGroup that is not within other tag groups. The four nodes that currently have this attribute are: Definition, Event-context, Onset, Offset. The consequence is that these tag groups cannot be nested. For example, Definitions cannot be nested. This is because the Definition syntax is: (Definition/defname, ( actual definition tags)) Right now topLevelTagGroup is being handled directly from validation using the schema, but the part that there can't be any other tags besides the Definition and its tag group is hard-coded in the validator. In a future round of schema revisions we will add another schema attribute to enforce this, since there will be some other tags such as Event-context that fall into this category. Since we aren't handling anything about Event-context yet except for its straight schema validation, we didn't do it for the first round. We don't want Onset and Offset in the same tag group either. This is currently hard-coded, but I think it can be taken care of by a single schema attribute.

…

On Fri, Aug 20, 2021 at 3:03 AM MichaelJachan ***@***.***> wrote: 1. I do not understand the attributes "tagGroup" and "topLevelTagGroup". Can u halp me?:) — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#323 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAJCJOWLEHD42O7UHPI444TT5YD35ANCNFSM5CLZFH3Q> . Triage notifications on the go with GitHub Mobile for iOS <https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android <https://play.google.com/store/apps/details?id=com.github.android&utm_campaign=notification-email> .

[VisLab]

We have code in Python to do it from github. What did you have in mind? We have a bunch of web services which are accessible in matlab. These are available with examples of how to call them in the webtools/examples subdirectory of hed-standard/hed-python on github. All of these services are callable from the server https://hedtools.ucsd.edu/hed. If it would be helpful to have a web service to get the XML callable from MATLAB, let me know. I can easily put it in for the next deployment release of the tools.

…

On Wed, Aug 18, 2021 at 9:06 AM MichaelJachan ***@***.***> wrote: So, your XML-source, which i can download in the end from https://github.com/hed-standard/hed-specification/blob/master/hedxml/HED8.0.0.xml, is generated by some Mediawiki-Software, i understand. Is there maybe another way for me to download the xml? — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#323 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAJCJOTNFN6DCMFTEYERESTT5O4Y7ANCNFSM5CLZFH3Q> . Triage notifications on the go with GitHub Mobile for iOS <https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android <https://play.google.com/store/apps/details?id=com.github.android&utm_campaign=notification-email> .

[VisLab]

Yes, this has been a languishing issue for a while (#19). Now that we are getting around to value classes, I'll escalate it to the proposed so it gets some attention.

…

On Thu, Aug 19, 2021 at 2:43 AM MichaelJachan ***@***.***> wrote: 1. IDEA: Would you like to consider providing separate numeric types for int and double (ie, float). For now, I am not able to distinguish btw. int and double, when I read your xml. Thank you :) — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#323 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAJCJOVONFE5E4SHNPKXU3TT5SY2TANCNFSM5CLZFH3Q> . Triage notifications on the go with GitHub Mobile for iOS <https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android <https://play.google.com/store/apps/details?id=com.github.android&utm_campaign=notification-email> .

[VisLab]

Could you post this as an issue on Github hed-standard/hed-matlab so we have a record of it being reported and fixed... Thanks! https://github.com/hed-standard/hed-matlab/issues

…

On Mon, Aug 23, 2021 at 8:48 AM MichaelJachan ***@***.***> wrote: 1. Warning in HED-MATLAB-TOOLS: WARNING: for plugin "HEDTools" version in the folder name "3.1.0" and in the eegplugin_ file "3.0.0" differ — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#323 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAJCJOQTT2I3NWKOWVCZADTT6JGTVANCNFSM5CLZFH3Q> . Triage notifications on the go with GitHub Mobile for iOS <https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android <https://play.google.com/store/apps/details?id=com.github.android&utm_campaign=notification-email> .