Editorial: Eschew vars outside of algorithms

Part of webmachinelearning#783

docs/SpecCodingConventions.md

@@ -95,6 +95,7 @@ Example:
 * When concisely defining a tensor's layout, use the syntax `*[ ... ]*` (e.g. _"nchw" means the input tensor has the layout *[batches, inputChannels, height, width]*_)
 * Format explanatory expressions using backticks, e.g. `` `max(0, x) + alpha * (exp(min(0, x)) - 1)` ``
 * In Web IDL `<pre class=idl>` blocks, wrap long lines to avoid horizontal scrollbars. 88 characters seems to be the magic number.
+* Avoid `<var>v</var>` or `|v|` outside of algorithms; Bikeshed interprets these as global variables which can mask errors. Just use `*v*`.
 
 
 ### Algorithms

index.bs

@@ -1767,7 +1767,7 @@ partial dictionary MLOpSupportLimits {
         - <dfn>axis</dfn>: The dimension to reduce. The value must be in the range [0, N-1] where N is the [=MLOperand/rank=] of the input tensor.
         - <dfn>options</dfn>: an optional {{MLArgMinMaxOptions}}. The optional parameters of the operation.
 
-    **Returns:** an {{MLOperand}}. The N-D tensor of the reduced shape. The values must be of type |options|.{{MLArgMinMaxOptions/outputDataType}} in the range [0, N-1] where N is the size of the input dimension specified by axis.
+    **Returns:** an {{MLOperand}}. The N-D tensor of the reduced shape. The values must be of type *options*.{{MLArgMinMaxOptions/outputDataType}} in the range [0, N-1] where N is the size of the input dimension specified by axis.
 </div>
 
 <table id=constraints-argMin-argMax class='data' link-for="MLGraphBuilder/argMin(input, axis, options), MLGraphBuilder/argMax(input, axis, options)">
@@ -4075,13 +4075,13 @@ partial dictionary MLOpSupportLimits {
 <div dfn-for="MLGraphBuilder/gru(input, weight, recurrentWeight, steps, hiddenSize, options)" dfn-type=argument>
     **Arguments:**
         - <dfn>input</dfn>: an {{MLOperand}}. The input 3-D tensor of shape *[steps, batchSize, inputSize]*.
-        - <dfn>weight</dfn>: an {{MLOperand}}. The 3-D input weight tensor of shape *[numDirections, 3 * hiddenSize, inputSize]*. The ordering of the weight vectors in the second dimension of the tensor shape is specified according to |options|.{{MLGruOptions/layout}}.
-        - <dfn>recurrentWeight</dfn>: an {{MLOperand}}. The 3-D recurrent weight tensor of shape *[numDirections, 3 * hiddenSize, hiddenSize]*. The ordering of the weight vectors in the second dimension of the tensor shape is specified according to |options|.{{MLGruOptions/layout}}.
+        - <dfn>weight</dfn>: an {{MLOperand}}. The 3-D input weight tensor of shape *[numDirections, 3 * hiddenSize, inputSize]*. The ordering of the weight vectors in the second dimension of the tensor shape is specified according to *options*.{{MLGruOptions/layout}}.
+        - <dfn>recurrentWeight</dfn>: an {{MLOperand}}. The 3-D recurrent weight tensor of shape *[numDirections, 3 * hiddenSize, hiddenSize]*. The ordering of the weight vectors in the second dimension of the tensor shape is specified according to *options*.{{MLGruOptions/layout}}.
         - <dfn>steps</dfn>: an {{unsigned long}} scalar. The number of time steps in the recurrent network. The value must be greater than 0.
         - <dfn>hiddenSize</dfn>: an {{unsigned long}} scalar. The value of the third dimension of the cell output tensor shape. It indicates the number of features in the hidden state.
         - <dfn>options</dfn>: an optional {{MLGruOptions}}. The optional parameters of the operation.
 
-    **Returns:** [=sequence=]<{{MLOperand}}>. The first element is a 3-D tensor of shape *[numDirections, batchSize, hiddenSize]*, the cell output from the last time step of the network. Additionally, if |options|.{{MLGruOptions/returnSequence}} is set to true, the second element is the 4-D output tensor of shape *[steps, numDirections, batchSize, hiddenSize]* containing every cell outputs from each time step in the temporal sequence.
+    **Returns:** [=sequence=]<{{MLOperand}}>. The first element is a 3-D tensor of shape *[numDirections, batchSize, hiddenSize]*, the cell output from the last time step of the network. Additionally, if *options*.{{MLGruOptions/returnSequence}} is set to true, the second element is the 4-D output tensor of shape *[steps, numDirections, batchSize, hiddenSize]* containing every cell outputs from each time step in the temporal sequence.
 </div>
 
 <table id=constraints-gru class='data' link-for="MLGraphBuilder/gru(input, weight, recurrentWeight, steps, hiddenSize, options)">
@@ -4379,8 +4379,8 @@ partial dictionary MLOpSupportLimits {
 <div dfn-for="MLGraphBuilder/gruCell(input, weight, recurrentWeight, hiddenState, hiddenSize, options)" dfn-type=argument>
     **Arguments:**
         - <dfn>input</dfn>: an {{MLOperand}}. The input 2-D tensor of shape *[batchSize, inputSize]*.
-        - <dfn>weight</dfn>: an {{MLOperand}}. The 2-D input weight tensor of shape *[3 * hiddenSize, inputSize]*. The ordering of the weight vectors in the first dimension of the tensor shape is specified according to *options.layout*.
-        - <dfn>recurrentWeight</dfn>: an {{MLOperand}}. The 2-D recurrent weight tensor of shape *[3 * hiddenSize, hiddenSize]*. The ordering of the weight vectors in the first dimension of the tensor shape is specified according to *options.layout*.
+        - <dfn>weight</dfn>: an {{MLOperand}}. The 2-D input weight tensor of shape *[3 * hiddenSize, inputSize]*. The ordering of the weight vectors in the first dimension of the tensor shape is specified according to *options*.{{MLGruCellOptions/layout}}.
+        - <dfn>recurrentWeight</dfn>: an {{MLOperand}}. The 2-D recurrent weight tensor of shape *[3 * hiddenSize, hiddenSize]*. The ordering of the weight vectors in the first dimension of the tensor shape is specified according to *options*.{{MLGruCellOptions/layout}}.
         - <dfn>hiddenState</dfn>: an {{MLOperand}}. The 2-D input hidden state tensor of shape *[batchSize, hiddenSize]*.
         - <dfn>hiddenSize</dfn>: an {{unsigned long}} scalar. The value of the second dimension of the output tensor shape. It indicates the number of features in the hidden state.
         - <dfn>options</dfn>: an optional {{MLGruCellOptions}}. The optional parameters of the operation.
@@ -4816,11 +4816,11 @@ partial dictionary MLOpSupportLimits {
 <dl dfn-type=dict-member dfn-for=MLInstanceNormalizationOptions>
     : <dfn>scale</dfn>
     ::
-        The 1-D tensor of the scaling values whose [=list/size=] is equal to the number of channels, i.e. the size of the feature dimension of the input. For example, for an |input| tensor with {{MLInputOperandLayout/"nchw"}} layout, the [=list/size=] is equal to |input|'s [=MLOperand/shape=][1].
+        The 1-D tensor of the scaling values whose [=list/size=] is equal to the number of channels, i.e. the size of the feature dimension of the input. For example, for an *input* tensor with {{MLInputOperandLayout/"nchw"}} layout, the [=list/size=] is equal to *input*'s [=MLOperand/shape=][1].
 
     : <dfn>bias</dfn>
     ::
-        The 1-D tensor of the bias values whose [=list/size=] is equal to the size of the feature dimension of the input. For example, for an |input| tensor with {{MLInputOperandLayout/"nchw"}} layout, the [=list/size=] is equal to |input|'s [=MLOperand/shape=][1].
+        The 1-D tensor of the bias values whose [=list/size=] is equal to the size of the feature dimension of the input. For example, for an *input* tensor with {{MLInputOperandLayout/"nchw"}} layout, the [=list/size=] is equal to *input*'s [=MLOperand/shape=][1].
 
     : <dfn>epsilon</dfn>
     ::
@@ -4971,11 +4971,11 @@ partial dictionary MLOpSupportLimits {
 <dl dfn-type=dict-member dfn-for=MLLayerNormalizationOptions>
     : <dfn>scale</dfn>
     ::
-        The N-D tensor of the scaling values whose shape is determined by the |axes| member in that each value in |axes| indicates the dimension of the input tensor with scaling values. For example, for an |axes| values of [1,2,3], the shape of this tensor is the list of the corresponding sizes of the input dimension 1, 2 and 3. When this member is not present, the scaling value is assumed to be 1.
+        The N-D tensor of the scaling values whose shape is determined by the {{MLLayerNormalizationOptions/axes}} member in that each value in {{MLLayerNormalizationOptions/axes}} indicates the dimension of the input tensor with scaling values. For example, for an {{MLLayerNormalizationOptions/axes}} values of [1,2,3], the shape of this tensor is the list of the corresponding sizes of the input dimension 1, 2 and 3. When this member is not present, the scaling value is assumed to be 1.
 
     : <dfn>bias</dfn>
     ::
-        The N-D tensor of the bias values whose shape is determined by the |axes| member in that each value in |axes| indicates the dimension of the input tensor with bias values. For example, for an |axes| values of [1,2,3], the shape of this tensor is the list of the corresponding sizes of the input dimension 1, 2 and 3. When this member is not present, the bias value is assumed to be 0.
+        The N-D tensor of the bias values whose shape is determined by the {{MLLayerNormalizationOptions/axes}} member in that each value in {{MLLayerNormalizationOptions/axes}} indicates the dimension of the input tensor with bias values. For example, for an {{MLLayerNormalizationOptions/axes}} values of [1,2,3], the shape of this tensor is the list of the corresponding sizes of the input dimension 1, 2 and 3. When this member is not present, the bias value is assumed to be 0.
 
     : <dfn>axes</dfn>
     ::
@@ -5376,13 +5376,13 @@ partial dictionary MLOpSupportLimits {
 <div dfn-for="MLGraphBuilder/lstm(input, weight, recurrentWeight, steps, hiddenSize, options)" dfn-type=argument>
     **Arguments:**
         - <dfn>input</dfn>: an {{MLOperand}}. The input 3-D tensor of shape *[steps, batchSize, inputSize]*.
-        - <dfn>weight</dfn>: an {{MLOperand}}. The 3-D input weight tensor of shape *[numDirections, 4 * hiddenSize, inputSize]*. The ordering of the weight vectors in the second dimension of the tensor shape is specified according to the |options|.{{MLLstmOptions/layout}}.
-        - <dfn>recurrentWeight</dfn>: an {{MLOperand}}. The 3-D recurrent weight tensor of shape *[numDirections, 4 * hiddenSize, hiddenSize]*. The ordering of the weight vectors in the second dimension of the tensor shape is specified according to |options|.{{MLLstmOptions/layout}}.
+        - <dfn>weight</dfn>: an {{MLOperand}}. The 3-D input weight tensor of shape *[numDirections, 4 * hiddenSize, inputSize]*. The ordering of the weight vectors in the second dimension of the tensor shape is specified according to *options*.{{MLLstmOptions/layout}}.
+        - <dfn>recurrentWeight</dfn>: an {{MLOperand}}. The 3-D recurrent weight tensor of shape *[numDirections, 4 * hiddenSize, hiddenSize]*. The ordering of the weight vectors in the second dimension of the tensor shape is specified according to *options*.{{MLLstmOptions/layout}}.
         - <dfn>steps</dfn>: an {{unsigned long}} scalar. The number of time steps in the recurrent network. The value must be greater than 0.
         - <dfn>hiddenSize</dfn>: an {{unsigned long}} scalar. The value of the third dimension of the cell output tensor shape. It indicates the number of features in the hidden state.
         - <dfn>options</dfn>: an optional {{MLLstmOptions}}. The optional parameters of the operation.
 
-    **Returns:** [=sequence=]<{{MLOperand}}>. The first element is a 3-D tensor of shape *[numDirections, batchSize, hiddenSize]*, the output hidden state from the last time step of the network. The second element is a 3-D tensor of shape *[numDirections, batchSize, hiddenSize]*, the output cell state from the last time step of the network. Additionally, if |options|.{{MLLstmOptions/returnSequence}} is set to true, the third element is the 4-D output tensor of shape *[steps, numDirections, batchSize, hiddenSize]* containing every output from each time step in the temporal sequence.
+    **Returns:** [=sequence=]<{{MLOperand}}>. The first element is a 3-D tensor of shape *[numDirections, batchSize, hiddenSize]*, the output hidden state from the last time step of the network. The second element is a 3-D tensor of shape *[numDirections, batchSize, hiddenSize]*, the output cell state from the last time step of the network. Additionally, if *options*.{{MLLstmOptions/returnSequence}} is set to true, the third element is the 4-D output tensor of shape *[steps, numDirections, batchSize, hiddenSize]* containing every output from each time step in the temporal sequence.
 </div>
 
 <table id=constraints-lstm class='data' link-for="MLGraphBuilder/lstm(input, weight, recurrentWeight, steps, hiddenSize, options)">
@@ -5735,8 +5735,8 @@ partial dictionary MLOpSupportLimits {
 <div dfn-for="MLGraphBuilder/lstmCell(input, weight, recurrentWeight, hiddenState, cellState, hiddenSize, options)" dfn-type=argument>
     **Arguments:**
         - <dfn>input</dfn>: an {{MLOperand}}. The input 2-D tensor of shape *[batchSize, inputSize]*.
-        - <dfn>weight</dfn>: an {{MLOperand}}. The 2-D input weight tensor of shape *[4 * hiddenSize, inputSize]*. The ordering of the weight vectors in the first dimension of the tensor shape is specified according to *options.layout*.
-        - <dfn>recurrentWeight</dfn>: an {{MLOperand}}. The 2-D recurrent weight tensor of shape *[4 * hiddenSize, hiddenSize]*. The ordering of the weight vectors in the first dimension of the tensor shape is specified according to *options.layout*.
+        - <dfn>weight</dfn>: an {{MLOperand}}. The 2-D input weight tensor of shape *[4 * hiddenSize, inputSize]*. The ordering of the weight vectors in the first dimension of the tensor shape is specified according to *options*.{{MLLstmCellOptions/layout}}.
+        - <dfn>recurrentWeight</dfn>: an {{MLOperand}}. The 2-D recurrent weight tensor of shape *[4 * hiddenSize, hiddenSize]*. The ordering of the weight vectors in the first dimension of the tensor shape is specified according to *options*.{{MLLstmCellOptions/layout}}.
         - <dfn>hiddenState</dfn>: an {{MLOperand}}. The 2-D input hidden state tensor of shape *[batchSize, hiddenSize]*.
         - <dfn>cellState</dfn>: an {{MLOperand}}. The 2-D input cell state tensor of shape *[batchSize, hiddenSize]*.
         - <dfn>hiddenSize</dfn>: an {{unsigned long}} scalar. The value of the second dimension of the output tensor shape. It indicates the number of features in the hidden state.

tools/lint.mjs

@@ -251,6 +251,13 @@ for (const algorithm of root.querySelectorAll('.algorithm')) {
   }
 }
 
+// Eschew vars outside of algorithms.
+const algorithmVars = new Set(root.querySelectorAll('.algorithm var'));
+for (const v of root.querySelectorAll('var').filter(v => !algorithmVars.has(v))) {
+  error(`Variable outside of algorithm: ${v.innerText}`);
+}
+
+
 // Prevent accidental normative references to other specs. This reports an error
 // if there is a normative reference to any spec *other* than these ones. This
 // helps avoid an autolink like [=object=] adding an unexpected reference to